API reference

API Documentation

One API for the models, agents, workflows, tasks and integrations in your MultiverseAI account. Requests run as the key's owner, with the same data and connected accounts you see in the app.

Base URL

https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1

Endpoint paths below are written as /v1/…; append them to https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api. GET /v1 returns an index of every endpoint.

OpenAI-compatible

Use the official openai client: change the base URL and API key

Every model

The chat, image and video models from the MultiverseAI app, listed at /v1/models

Agent tools

tools: true gives a model the full MultiverseAI toolkit, run server-side as you

Streaming

Server-sent events for chat, agents and conversations

Per-key switches

Turn workflows, agents and tasks access on or off for each key

Credits

AI calls are paid from your MultiverseAI credits; check them at /v1/credits

Quick start

// npm install openai
import OpenAI from 'openai';

const client = new OpenAI({
  apiKey: process.env.MULTIVERSE_API_KEY, // mverse_k_...
  baseURL: 'https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1',
});

const completion = await client.chat.completions.create({
  model: 'openrouter/openai/gpt-6-luna',
  messages: [{ role: 'user', content: 'Hello!' }],
});

console.log(completion.choices[0].message.content);

Authentication

Send a developer key in the Authorization header of every request. Keys start with mverse_k_ and are created in Settings → Developer. A signed-in user's Supabase access token is accepted too. Every request runs as the key's owner, so it can only see and change that account's data.

A key can spend your credits and act on your connected accounts. Keep it on the server — never ship it in client-side code or commit it to a repository.

Header

Authorization: Bearer mverse_k_YOUR_API_KEY

Example request

export MULTIVERSE_API_KEY=mverse_k_YOUR_API_KEY

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/credits" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

Key permissions

Each key has five switches. Turning one off blocks its endpoints with 403 FORBIDDEN; every other endpoint is open to any active key. Keys made before a switch existed keep access to its endpoints until you turn it off. Keys can't be created or revoked through the API.

Switch	Endpoints
`workflows`	/v1/workflows, /v1/executions, /v1/queue
`agents`	/v1/agents
`tasks`	/v1/tasks, /v1/boards
`posts`	/v1/posts, /v1/platforms
`tools`	/v1/tools

Need an API key?

Create one in Developer Settings

Limits, Credits & Lists

Rate limits

120 requests per minute per IP address and 300 requests per minute per API key. Over the limit, requests fail with 429 RATE_LIMITED and a retry_after_ms field in the error.

Credits

Model calls (chat, completions, images, videos, tool runs, agent and conversation replies) are paid from the key owner's MultiverseAI credits, the same as in the app. When the balance runs out they fail with 402 INSUFFICIENT_CREDITS. Check the balance with GET /v1/credits.

Lists and pagination

List endpoints take ?limit (default 50, max 100 unless noted) and ?offset.

{
  "object": "list",
  "data": [ … ],
  "total": 128,
  "limit": 50,
  "offset": 0
}

Streaming

Chat completions, agent chat and conversation messages accept stream: true and answer with server-sent events (text/event-stream): one data: {…} line per event, ending with data: [DONE]. Chat completion chunks follow the OpenAI format, so OpenAI-compatible clients stream them unchanged.

SDK & CLI

Today, call the API over plain HTTP or with any OpenAI-compatible client (see the quick start). A dedicated TypeScript SDK and a command-line tool are in development.

TypeScript SDK

Coming soon

@multiverseai/sdk is not published to npm yet. It can be built from the sdk/ folder of the MultiverseAI repository in the meantime.

CLI

Coming soon

multiverse-cli is not published to npm yet. To try it, build it from the multiverse-cli/ folder of the repository:

cd multiverse-cli
npm install && npm run build && npm link

Models

Every chat, image and video model the MultiverseAI app offers. Use a model's id as the "model" field of any request. These endpoints need no authentication.

GET

/v1/models

List available models

Text and image models: id, name, provider, modality, tier, description, context_length, tools (supports agent tools), cost_per_1k_tokens.
Video models add durations, default_duration, aspect_ratios, audio and usd_per_second.
The default model is marked "default": true (chat: openrouter/openai/gpt-6-luna).

Query parameters

Field	Type	Required	Description
`modality`	`string`	No	Filter by "text", "image" or "video"

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/models?modality=text"

GET

/v1/models/{id}

Get one model

Model ids contain slashes; append the id as-is, e.g. /v1/models/openrouter/anthropic/claude-opus-5.5. Unknown ids return 404 MODEL_NOT_FOUND.

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/models/openrouter/anthropic/claude-opus-5.5"

GET

/v1/health

Health check

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/health"

GET

/v1

Endpoint index

Returns the API name, version, documentation link and a map of every endpoint.

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1"

Chat Completions

OpenAI-compatible chat completions across every MultiverseAI model. Point the official openai client (or any OpenAI-compatible library) at the /v1 base URL and use your mverse_k_ key as the API key. Calls are paid from your MultiverseAI credits.

Roles: system (or developer), user and assistant. The last message must be from the user. Tool-role messages are not supported.
Image models answer with markdown image(s); video models start a video job (see Videos). Both are surfaced in the multiverse field of the response.
Image models use the whole conversation: a follow-up like "make it winter" revises the last image, and images in the last user message are used as references to work from (PNG, JPEG or WebP).
finish_reason is "length" when the reply was cut off at max_tokens (some providers also append a short note to the text). An unsupported model or malformed request is a 400, an oversized image a 413, and an empty model reply a 502 UPSTREAM_ERROR — 4xx errors are not worth retrying unchanged.

POST

/v1/chat/completions

Create a chat completion

Request body

Field	Type	Required	Description
`model`	`string`	No	Model id from GET /v1/modelsDefault: `openrouter/openai/gpt-6-luna`
`messages`	`array`	Yes	Objects with role and content. content is a string or an array of parts: { type: "text", text } and { type: "image_url", image_url: { url } }.
`stream`	`boolean`	No	Stream chat.completion.chunk events over SSEDefault: `false`
`stream_options`	`object`	No	{ include_usage: true } adds a final usage chunk when streaming
`temperature`	`number`	No	Sampling temperature, 0–2
`max_tokens`	`integer`	No	Maximum tokens to generate (max_completion_tokens is accepted too)
`top_p`	`number`	No	Nucleus sampling, 0–1
`frequency_penalty`	`number`	No	-2 to 2
`presence_penalty`	`number`	No	-2 to 2
`tools`	`true \| string[]`	No	true enables every MultiverseAI agent tool; an array enables only the named tools (names from GET /v1/tools). Tools run server-side as you. OpenAI-style function definitions are rejected with UNSUPPORTED_TOOLS.
`max_tool_iterations`	`integer`	No	Cap on tool-call rounds when tools are enabled (1–20)

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/chat/completions" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "openrouter/anthropic/claude-opus-5.5",
  "messages": [
    {
      "role": "system",
      "content": "You are a concise assistant."
    },
    {
      "role": "user",
      "content": "Explain vector databases in two sentences."
    }
  ],
  "max_tokens": 300
}'

POST

/v1/chat/completions

Streaming (stream: true)

The response is text/event-stream: "data: {chat.completion.chunk}" lines, ending with "data: [DONE]".
With tools enabled, tool activity arrives as extra chunks with choices: [] and multiverse: { event, tool_call } (event is tool_call_start, tool_call_result or tool_call_error).
A mid-stream failure arrives as a chunk with choices: [] and error: { message, code }.

curl -X POST -N "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/chat/completions" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "messages": [
    {
      "role": "user",
      "content": "Write a haiku about the ocean."
    }
  ],
  "stream": true,
  "stream_options": {
    "include_usage": true
  }
}'

POST

/v1/chat/completions

With MultiverseAI agent tools (tools: true)

tools: true gives the model every MultiverseAI agent tool (web search, scraping, workflows, tasks, integrations…); pass an array such as ["web_search"] to limit it.
The tools run on the server as the key's owner, so they see your connected accounts and data. The final answer comes back as a normal completion, with each tool call listed under multiverse.tool_calls.

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/chat/completions" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "messages": [
    {
      "role": "user",
      "content": "What are the top AI headlines today? Cite sources."
    }
  ],
  "tools": [
    "web_search"
  ],
  "max_tool_iterations": 4
}'

POST

/v1/chat/completions

With image input

Images are accepted only in the last user message (up to 10). Each url must be https:// or a data: URL; JPEG, PNG, GIF or WebP, 20 MB max.
With an image model the images are references for the new image (up to 4 are used; GIF is not accepted there).

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/chat/completions" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "openrouter/anthropic/claude-opus-5.5",
  "messages": [
    {
      "role": "user",
      "content": [
        {
          "type": "text",
          "text": "What is in this picture?"
        },
        {
          "type": "image_url",
          "image_url": {
            "url": "https://example.com/photo.jpg"
          }
        }
      ]
    }
  ]
}'

Completions

Legacy single-prompt completions: the prompt is sent as one user message. Prefer Chat Completions for new code.

POST

/v1/completions

Create a text completion

When an image model is used the response has the Images shape: { created, model, data: [{ url }] }.

Request body

Field	Type	Required	Description
`prompt`	`string`	Yes	The prompt
`model`	`string`	No	Model id from GET /v1/modelsDefault: `openrouter/openai/gpt-6-luna`
`modality`	`string`	No	"image" generates images instead (with gpt-image-1.5 unless model is an image model). Passing an image model does the same.
`temperature`	`number`	No	Sampling temperature, 0–2
`max_tokens`	`integer`	No	Maximum tokens to generate (max_completion_tokens is accepted too)
`top_p`	`number`	No	Nucleus sampling, 0–1
`frequency_penalty`	`number`	No	-2 to 2
`presence_penalty`	`number`	No	-2 to 2
`tools`	`true \| string[]`	No	true enables every MultiverseAI agent tool; an array enables only the named tools (names from GET /v1/tools). Tools run server-side as you. OpenAI-style function definitions are rejected with UNSUPPORTED_TOOLS.
`max_tool_iterations`	`integer`	No	Cap on tool-call rounds when tools are enabled (1–20)

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/completions" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "prompt": "Write a tagline for a hiking app.",
  "max_tokens": 60,
  "temperature": 0.8
}'

Images

Generate images with any image model (GET /v1/models?modality=image). Images are stored in MultiverseAI and returned as URLs.

POST

/v1/images/generations

Generate images

If some of several images fail, the ones that succeeded are returned with a warning field.
With image, the model edits or builds on the reference ("the full landing page this hero belongs to", "same scene in winter"). GPT Image, Nano Banana, Seedream and FLUX.2 take references; DALL·E 3 and Ideogram return 400. GPT Image keeps the reference's aspect ratio.

Request body

Field	Type	Required	Description
`prompt`	`string`	Yes	Description of the image
`model`	`string`	No	An image model id; a non-image model returns 400 INVALID_MODEL_FOR_MODALITYDefault: `gpt-image-1.5`
`n`	`integer`	No	Number of images, 1–4Default: `1`
`image`	`string \| string[]`	No	Reference image(s) to work from: an https:// or data: URL, or an array of up to 4. PNG, JPEG or WebP, 20 MB max each. The prompt then says what to keep and what to change.

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/images/generations" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "prompt": "A watercolor lighthouse at dawn",
  "n": 1
}'

Videos

Asynchronous video generation. Submit a job, then poll it until status is completed (video_url is set) or failed. Cost depends on the model and duration and is charged from your credits.

Job status: queued → running → saving → completed, or failed / cancelled.

POST

/v1/videos

Start a video job

Request body

Field	Type	Required	Description
`prompt`	`string`	Yes	What the video should show
`model`	`string`	No	A video model id (GET /v1/videos/models)Default: `video/fal/fal-ai/veo3.1/fast`
`duration`	`integer`	No	Seconds; snapped to the model's closest allowed duration
`aspect_ratio`	`string`	No	e.g. "16:9" or "9:16", if the model supports it
`image_url`	`string`	No	Optional first frame (image-to-video): an https:// or data:image/… URL

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/videos" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "prompt": "A drone shot over a misty pine forest at sunrise",
  "duration": 6,
  "aspect_ratio": "16:9"
}'

GET

/v1/videos/{id}

Get a video job (refreshes its status from the provider)

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/videos/9b2f6c1e-4d7a-4e3b-a5c8-1f0e2d3c4b5a" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

GET

/v1/videos

List your video jobs

Query parameters

Field	Type	Required	Description
`limit`	`integer`	No	Up to 100Default: `30`

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/videos?limit=10" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

GET

/v1/videos/models

List video models

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/videos/models" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

DELETE

/v1/videos/{id}

Cancel a video job

curl -X DELETE "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/videos/9b2f6c1e-4d7a-4e3b-a5c8-1f0e2d3c4b5a" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

Avatar Videos (HeyGen & Tavus)

AI avatars and replicas that speak your script, video translation, and live AI video calls. Avatar videos are jobs in the same queue as other videos: start one with POST /v1/videos and a heygen/… or tavus/… model, then poll GET /v1/videos/{id} until status is completed (video_url is set). They are charged in credits by the finished length — or free when you have saved your own HeyGen / Tavus API key in Settings → Integrations.

Prices (USD per finished minute, converted to credits): HeyGen Avatar IV $4.83 ($2.31 for photo avatars), Avatar V $7.20, Avatar III $0.60–0.99; HeyGen translation $0.81 (precision $1.50, audio-only $0.57); Tavus replica video $1.00; Tavus live conversation $0.37 per minute the replica is in the call.
Your balance must cover the estimate up front (about 2.5 spoken words per second; 60 seconds when you give an audio file or a video to translate). Failed and cancelled jobs are free.
Renders usually take 2–10 minutes. Finished videos are copied to MultiverseAI storage (up to 50 MB; larger ones keep the provider link, which expires).

POST

/v1/videos

HeyGen avatar video (model heygen/avatar)

Request body

Field	Type	Required	Description
`model`	`string`	Yes	"heygen/avatar"
`avatar_id`	`string`	Yes	An avatar look id from GET /v1/videos/avatars?provider=heygen
`script`	`string`	No	What the avatar says (up to 5,000 characters). Give this or audio_url.
`voice_id`	`string`	No	Voice id (GET /v1/videos/avatars?provider=heygen&kind=voices); default: the avatar's voice
`audio_url`	`string`	No	Public https audio to lip-sync instead of a script
`engine`	`string`	No	avatar_iv, avatar_v or avatar_iii (must be in the avatar's supported_engines)Default: `avatar_iv`
`aspect_ratio`	`string`	No	16:9, 9:16, 1:1, 4:5, 5:4 or autoDefault: `16:9`
`resolution`	`string`	No	720p or 1080pDefault: `720p`
`background`	`string`	No	A hex color (#0f172a) or an https image URL
`motion_prompt`	`string`	No	Body motion and gestures (photo avatars, or engine avatar_v)
`language`	`string`	No	Voice locale, e.g. en-US
`title`	`string`	No	Name in your HeyGen dashboard

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/videos" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "heygen/avatar",
  "avatar_id": "Abigail_expressive_2024112501",
  "script": "Hi! Thanks for signing up — here is how to get started.",
  "aspect_ratio": "9:16"
}'

POST

/v1/videos

HeyGen video translation (model heygen/translate)

Request body

Field	Type	Required	Description
`model`	`string`	Yes	"heygen/translate"
`video_url`	`string`	Yes	Public https URL of the video to translate
`output_language`	`string`	Yes	Target language name, e.g. "Spanish (Spain)" (GET /v1/videos/avatars?provider=heygen&kind=languages)
`mode`	`string`	No	speed, or precision for better lip-syncDefault: `speed`
`translate_audio_only`	`boolean`	No	Dub the audio only, without lip-sync (cheapest)Default: `false`
`input_language`	`string`	No	Source language (detected when omitted)
`speaker_num`	`integer`	No	Number of speakers (improves separation)
`title`	`string`	No	Name in your HeyGen dashboard

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/videos" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "heygen/translate",
  "video_url": "https://example.com/launch.mp4",
  "output_language": "Spanish (Spain)"
}'

POST

/v1/videos

Tavus replica video (model tavus/replica)

Request body

Field	Type	Required	Description
`model`	`string`	Yes	"tavus/replica"
`replica_id`	`string`	Yes	A replica id from GET /v1/videos/avatars?provider=tavus
`script`	`string`	No	What the replica says. Give this or audio_url.
`audio_url`	`string`	No	Public https .mp3 / .wav to lip-sync instead of a script
`video_name`	`string`	No	A name for the video
`background_url`	`string`	No	A public website scrolled behind the replica
`background_source_url`	`string`	No	A public https video to use as the background
`fast`	`boolean`	No	Faster render without backgrounds or thumbnailsDefault: `false`

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/videos" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "tavus/replica",
  "replica_id": "r9fa0878977a",
  "script": "Hey Sam, I took a look at your site…",
  "background_url": "https://example.com"
}'

GET

/v1/videos/avatars

List avatars, voices, languages, replicas or personas (free)

Query parameters

Field	Type	Required	Description
`provider`	`string`	No	heygen or tavusDefault: `heygen`
`kind`	`string`	No	heygen: avatars (default), voices, languages. tavus: replicas (default), personas.
`avatar_type`	`string`	No	heygen avatars: studio_avatar, digital_twin or photo_avatar
`ownership`	`string`	No	heygen avatars: public (stock) or private (yours)
`language`	`string`	No	heygen voices: filter by language
`gender`	`string`	No	heygen voices: filter by gender
`limit`	`integer`	No	Page size (heygen avatars ≤ 50, voices ≤ 100)
`token`	`string`	No	heygen: next_token from the previous page
`page`	`integer`	No	tavus: page number

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/videos/avatars?provider=heygen&ownership=public&limit=2" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

POST

/v1/videos/conversations

Start a live AI video call with a Tavus replica

Open conversation_url in a browser to join. The call ends by itself at max_minutes, or shortly after everyone leaves.
Charged $0.37 per minute the replica is in the call (in 6-second steps), when it ends. On credits, max_minutes is also limited to what your balance covers.

Request body

Field	Type	Required	Description
`replica_id`	`string`	Yes	Replica id (optional when persona_id has a default replica)
`persona_id`	`string`	No	A persona (GET /v1/videos/avatars?provider=tavus&kind=personas): its LLM, prompt and tools
`conversational_context`	`string`	No	What the call is about / instructions for the replica
`custom_greeting`	`string`	No	The first thing the replica says
`conversation_name`	`string`	No	A name
`language`	`string`	No	e.g. english or multilingual
`callback_url`	`string`	No	Tavus posts conversation events (transcript, shutdown) here
`max_minutes`	`integer`	No	Hard limit on the call, 1–30Default: `10`

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/videos/conversations" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "replica_id": "r9fa0878977a",
  "conversational_context": "Help me practice a job interview for a product manager role.",
  "max_minutes": 15
}'

GET

/v1/videos/conversations/{id}

Get a conversation (status, billed seconds, charge)

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/videos/conversations/8f2a6c1d-3b4e-4f5a-8c7d-1e2f3a4b5c6d" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

POST

/v1/videos/conversations/{id}/end

End a conversation and settle its bill (also DELETE /v1/videos/conversations/{id})

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/videos/conversations/8f2a6c1d-3b4e-4f5a-8c7d-1e2f3a4b5c6d/end" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

GET

/v1/videos/conversations

List your conversations

Query parameters

Field	Type	Required	Description
`limit`	`integer`	No	Up to 100Default: `20`

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/videos/conversations?limit=10" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

Audio

Speech, transcription and sound effects by ElevenLabs. Generated audio is saved as an mp3 at a public, unguessable URL. Runs on your own ElevenLabs key if you saved one in Settings → Integrations (no MultiverseAI credits); otherwise it is charged from your credits at ElevenLabs' API prices. Listing voices is free.

Errors: 402 INSUFFICIENT_CREDITS when you (or, with your own key, your ElevenLabs account) are out of credits; 401 when your saved ElevenLabs key is rejected; 429 when ElevenLabs is rate limiting.

POST

/v1/audio/speech

Text to speech

Key permission: tools

Send Accept: audio/mpeg to get the mp3 bytes back instead of JSON (the URL is in the X-Audio-Url header).
Flash and Turbo models cost half as much per character as Multilingual v2, v3 and v4.

Request body

Field	Type	Required	Description
`input`	`string`	Yes	What to say: up to 10,000 characters (5,000 on eleven_v3)
`voice`	`string`	No	A voice_id from GET /v1/audio/voices, or alloy \| echo \| fable \| onyx \| nova \| shimmerDefault: `nova`
`model`	`string`	No	eleven_multilingual_v2 \| eleven_v3 \| eleven_v4 \| eleven_v4_turbo \| eleven_flash_v2_5 \| eleven_turbo_v2_5. tts-1 and tts-1-hd are accepted as aliases.Default: `eleven_multilingual_v2`
`response_format`	`string`	No	mp3, or a bitrate: mp3_44100_128 \| mp3_44100_192 \| mp3_44100_64 \| mp3_22050_32Default: `mp3`
`stability`	`number`	No	0–1; lower is more expressive
`similarity_boost`	`number`	No	0–1; how closely to match the voice

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/audio/speech" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "input": "Welcome back! Your report is ready.",
  "voice": "nova",
  "model": "eleven_flash_v2_5"
}'

POST

/v1/audio/transcriptions

Speech to text (ElevenLabs Scribe)

Key permission: tools

Charged per second of audio. File uploads are not supported yet; host the file and send its URL.

Request body

Field	Type	Required	Description
`audio_url`	`string`	Yes	Public https URL of the audio or video, up to 50 MB
`language`	`string`	No	ISO-639 code (e.g. en, spa) if known; detected otherwise
`diarize`	`boolean`	No	Label who is speakingDefault: `false`

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/audio/transcriptions" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "audio_url": "https://example.com/interview.mp3",
  "diarize": true
}'

GET

/v1/audio/voices

List voices (free)

Key permission: tools

With your own key, your cloned and library voices are included alongside the premade ones.

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/audio/voices" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

POST

/v1/audio/sound-effects

Generate a sound effect

Key permission: tools

Charged per second of audio generated.

Request body

Field	Type	Required	Description
`prompt`	`string`	Yes	Describe the sound (up to 1,000 characters)
`duration_seconds`	`number`	No	0.5–30; chosen by the model if omitted
`prompt_influence`	`number`	No	0–1; how literally to follow the promptDefault: `0.3`

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/audio/sound-effects" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "prompt": "Heavy wooden door creaking open in an empty hall",
  "duration_seconds": 4
}'

Meetings

AI meeting notetaker. Send a bot into a Google Meet, Zoom, Microsoft Teams or Webex call (now or scheduled); after the call you get a speaker-labelled transcript plus an AI summary, action items and decisions. Recordings are made with Recall.ai.

Status: scheduled → joining → in_waiting_room → in_call → recording → processing → done, or failed / cancelled. Someone in the meeting may have to admit the bot from the waiting room.
Notes are ready a few minutes after the call ends. Poll GET /v1/meetings/{id}, or start a workflow with a "Meeting ended" trigger.
Billing: bot time is charged from your credits at Recall.ai cost (~$0.65/hour, prorated per second) plus the standard markup, once the meeting ends. If you saved your own Recall.ai key (Settings → Integrations), bot time is billed by Recall.ai and not in credits. The AI notes use credits either way.
Limits: 3 bots in calls at once and 25 scheduled per account; a recording stops after 4 hours.

POST

/v1/meetings

Send the notetaker into a meeting

Returns 201 with the meeting (status joining or scheduled). 402 when out of credits, 409 TOO_MANY_BOTS at the limit, 400 INVALID_MEETING_URL for unsupported links.

Request body

Field	Type	Required	Description
`meeting_url`	`string`	Yes	meet.google.com/…, zoom.us/j/…, teams.microsoft.com/l/meetup-join/…, teams.live.com/meet/…, or *.webex.com/…
`title`	`string`	No	Name for the meeting in your notes
`bot_name`	`string`	No	Name the bot shows in the callDefault: `MultiverseAI Notetaker`
`join_at`	`string`	No	ISO 8601 time to join (schedule, up to 60 days ahead). Omit to join now.
`project_id`	`string`	No	File the meeting under a project (its id or your external_id)
`language`	`string`	No	Transcription language codeDefault: `auto`

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/meetings" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "meeting_url": "https://meet.google.com/abc-defg-hij",
  "title": "Weekly product sync"
}'

GET

/v1/meetings

List your meetings (newest first, without transcripts)

Query parameters

Field	Type	Required	Description
`limit`	`integer`	No	Up to 100Default: `50`
`offset`	`integer`	No	Default: `0`
`status`	`string`	No	Only this status (e.g. done)
`project_id`	`string`	No	Only meetings of this project
`search`	`string`	No	Search titles, summaries and transcripts

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/meetings?status=done&limit=10" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

GET

/v1/meetings/{id}

Get a meeting with its notes and transcript (refreshes the status while it is unfinished)

Query parameters

Field	Type	Required	Description
`include_transcript`	`boolean`	No	false leaves out transcript and transcript_textDefault: `true`

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/meetings/3f1c2b4a-8d7e-4c6b-9a5f-2e1d0c9b8a7f" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

PATCH

/v1/meetings/{id}

Rename a meeting

Request body

Field	Type	Required	Description
`title`	`string`	Yes	New title (empty clears it)

curl -X PATCH "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/meetings/3f1c2b4a-8d7e-4c6b-9a5f-2e1d0c9b8a7f" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "title": "Q4 roadmap review"
}'

POST

/v1/meetings/{id}/leave

Make the notetaker leave the call now (notes follow); a scheduled notetaker is cancelled

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/meetings/3f1c2b4a-8d7e-4c6b-9a5f-2e1d0c9b8a7f/leave" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{}'

POST

/v1/meetings/{id}/resummarize

Regenerate the AI summary, action items and decisions from the transcript (uses credits)

Only for meetings with status done and a transcript; 409 NOT_READY otherwise.

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/meetings/3f1c2b4a-8d7e-4c6b-9a5f-2e1d0c9b8a7f/resummarize" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{}'

DELETE

/v1/meetings/{id}

Delete a meeting, its notes and its recording and transcript at Recall.ai

409 MEETING_IN_PROGRESS while the bot is in the call; leave first. Deleting a scheduled meeting cancels its bot.

curl -X DELETE "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/meetings/3f1c2b4a-8d7e-4c6b-9a5f-2e1d0c9b8a7f" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

AngelSend

AngelSend email outreach (campaigns, sequences, per-recipient controls, senders and warm-up, opt-outs, the send queue, inbox, CRM, webhooks) on your own AngelSend account. All 148 AngelSend gateway actions are available with the same fields AngelSend documents; MultiverseAI adds the action catalog and "AngelSend event" workflow triggers.

Add your AngelSend API key (sk_…) in Settings → Integrations → AngelSend. Calls act as that AngelSend account; MultiverseAI credits are not used (AngelSend bills its own credits for AI actions).
Errors keep AngelSend's HTTP status — 400 invalid input or confirmation_required, 401 bad AngelSend key, 402 out of AngelSend credits, 403 missing key scope, 404, 409, 429, 5xx — as { error: { message: "AngelSend 404: …", code, status, details } }. A missing key is 400 ANGELSEND_NOT_CONFIGURED.
Destructive actions (delete-*, remove-*, stop-campaign-recipients, …) do nothing without "confirmed": true and return 400 CONFIRMATION_REQUIRED with a preview; "dry_run": true returns that preview with 200 (delete-campaign, delete-contact, delete-company, delete-deal, delete-chatbot, remove-contacts-from-campaign, remove-email-optout; plus the dry-run operations of manage-email-verification and manage-issuer-funnel). dry_run on any other action is refused with 400 rather than run for real.
create-full-campaign is retry-safe with an Idempotency-Key header (or "idempotency_key"): a repeat returns the first campaign with replayed: true. Campaigns created through an AngelSend key pause themselves if that key is revoked (pause_on_key_revoke, default true).
Developer keys need the Tools switch (Settings → Developer).

GET

/v1/angelsend/actions

The AngelSend action catalog

Key permission: tools

Query parameters

Field	Type	Required	Description
`group`	`string`	No	campaigns, recipients, senders, records, contacts, crm, compliance, webhooks, inbox, analytics, knowledge, chatbots, intelligence, quality, intent, booking, integrations, outreach, review, platform, issuer_funnel, core
`kind`	`string`	No	read \| compute (AI / paid) \| write \| destructive
`q`	`string`	No	Words to match

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/angelsend/actions?group=recipients" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

GET

/v1/angelsend/actions/{action}

One action's documentation and parameters

Key permission: tools

POST

/v1/angelsend/{action}

Run one AngelSend action — the body is the action's fields

Key permission: tools

Per-recipient control: pause-campaign-recipients / resume-campaign-recipients / stop-campaign-recipients { campaign_id, contact_ids | emails, reason }, list-campaign-recipients, set-recipient-merge-vars, set-recipient-start-step { step_number }.
Sequences and senders: update-campaign-step { campaign_id, step_number, operation: update | insert | delete, … }; get-sender-limits, update-sender-limits { sender_address_id, daily_limit ≤ 5000, warmup: { enabled, start_daily, daily_increment, restart } }.
Queue and compliance: list-campaign-queue rows carry bcc, compliance_archive, compliance_record, bounce and unsubscribe details; list-optouts pages with limit ≤ 1000, offset, since/until, campaign_id and order: "asc"; manage-compliance-archive { mode: "get" } includes team projects (owned: false).
Webhooks on your AngelSend account: register-outbound-webhook (with project_id), list-webhooks, update-webhook, delete-webhook, rotate-webhook-secret, test-webhook, list-webhook-deliveries, redeliver-webhook-delivery.

Request body

Field	Type	Required	Description
`…`	`any`	No	The action's own fields (see GET /v1/angelsend/actions/{action}). Multiplexed actions (manage-tags, manage-block-list, configure-follow-up-rule, …) take their operation as "action"; manage-compliance-archive as "mode".
`dry_run`	`boolean`	No	Preview a delete/removal without changing anything
`confirmed`	`boolean`	No	Required to run destructive actions
`idempotency_key`	`string`	No	create-full-campaign only (or the Idempotency-Key header)

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/angelsend/stop-campaign-recipients" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "campaign_id": "7c1f0a2e-5b3d-4e8f-9a6b-1d2c3e4f5a6b",
  "emails": [
    "sam@acme.com"
  ],
  "reason": "Replied on LinkedIn"
}'

POST

/v1/angelsend/connect

Send AngelSend events to your "AngelSend event" workflows

Key permission: tools

Registers a signed webhook on your AngelSend account pointing at MultiverseAI; the signing secret stays on our side. Deliveries are verified (X-AngelSend-Signature HMAC, 5-minute window), deduped on the event id and start each Active workflow whose Trigger is "AngelSend event" with matching filters, as you.
Trigger data: {{trigger.event}}, campaign_id, email, contact_id, message_id, subject, reply_content, bounce_type, reason, meeting_id, project_id, data.

Request body

Field	Type	Required	Description
`project_id`	`string`	No	Only this AngelSend project's events (including teammates' campaigns in it). Omit for account-wide.
`events`	`string[]`	No	Default: reply_received, bounced, unsubscribed, booking, recipient_stopped. Any AngelSend event name (message.opened, campaign.paused, …) or "*" also works.
`replace`	`boolean`	No	Re-register even if already connected

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/angelsend/connect" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{}'

GET

/v1/angelsend/connections

Your AngelSend webhook connections

Key permission: tools

Query parameters

Field	Type	Required	Description
`verify`	`boolean`	No	Also check that AngelSend still has each webhook (one is removed when the AngelSend key that registered it is revoked)

DELETE

/v1/angelsend/connections/{id}

Disconnect (deletes the webhook at AngelSend)

Key permission: tools

Tools

The MultiverseAI agent toolkit: your workspace (tasks, teams, contexts, memories, projects, credits), prompts and workflows (the Prompt Runner, queue, builder, schedules, templates), saved agents, social and messaging (X, Instagram, Facebook Pages and Groups, Threads, LinkedIn, TikTok, WhatsApp, Messenger, Telegram), ads and commerce, Google, GitHub and productivity apps, web research, and image/video generation. List the tools and their JSON-schema parameters, or run one directly as yourself — integration tools act on the accounts you have connected. The same tools power tools: true in Chat Completions and the MCP server. The full catalog is below.

GET

/v1/tools

List agent tools

Key permission: tools

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/tools" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

GET

/v1/tools/{name}

Get one tool

Key permission: tools

Unknown names return 404 UNKNOWN_TOOL.

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/tools/web_search" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

POST

/v1/tools/{name}

Run a tool

Key permission: tools

POST /v1/tools/{name}/run is an alias.

Request body

Field	Type	Required	Description
`arguments`	`object`	No	The tool's arguments, matching its parameters schemaDefault: `{}`
`confirm`	`boolean`	No	Required for tools with requires_approval: true (they change data or send messages), except the actions in their read_only_actions; without it the call returns 409 CONFIRMATION_REQUIRED
`project_id`	`string`	No	For tools that take project_id (facebook_ads, social_post_scheduler): run on that project's connections. Same as arguments.project_id (which wins). Other tools return 400.

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/tools/web_search" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "arguments": {
    "query": "MultiverseAI"
  }
}'

POST

/v1/tools/facebook_ads

Run Meta ads for a project (or your own Meta account)

Key permission: tools

Read-only (no confirm): list_ad_accounts, list_pages, list_pixels (ad_account_id), list_campaigns (ad_account_id), get_campaign (campaign_id; includes its ad sets and ads), list_adsets, list_ads, list_creatives, list_audiences, get_insights.
get_insights: the object is ad_id, adset_id, campaign_id or ad_account_id (first given); time_range { since, until } (YYYY-MM-DD) or date_preset (default last_7d); time_increment 1 for one row per day; level account|campaign|adset|ad; breakdowns; fields; limit / after.
Writes (confirm: true): create_campaign (ad_account_id, name, objective, daily_budget | lifetime_budget, special_ad_categories), create_adset (ad_account_id, campaign_id, name, targeting | country, optimization_goal, billing_event, promoted_object, budgets, start_time, end_time), upload_image (ad_account_id, image_url → image_hash), create_ad_creative (ad_account_id, page_id, link_url, message, headline, description, image_hash | image_url | video_id, call_to_action, instagram_user_id), create_ad (ad_account_id, adset_id, creative_id, name), create_audience, update (campaign_id | adset_id | ad_id with status and/or daily_budget, lifetime_budget, name), pause, resume, delete (also needs arguments.confirm: true).
Campaigns, ad sets and ads are created PAUSED unless status is ACTIVE. Budgets and bid amounts are integers in the ad account currency's minor units (5000 = $50.00). List results page with paging.cursors.after → after.
data.connection says which token ran the call: { source: "project", project_id, account_id, meta_user_id } or { source: "user" }.

Request body

Field	Type	Required	Description
`project_id`	`string`	No	Project id or external_id whose meta_ads connection to use (POST /v1/projects/{id}/connect with platform meta_ads). Without it: your own Meta connection. A project with no meta_ads connection is an error — it never falls back to your ad account.
`arguments`	`object`	Yes	action plus its fields (see below and GET /v1/tools/facebook_ads)
`confirm`	`boolean`	No	Needed for every action except the read-only ones

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/tools/facebook_ads" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "project_id": "acme-client-42",
  "arguments": {
    "action": "get_insights",
    "ad_account_id": "act_123",
    "level": "campaign",
    "time_range": {
      "since": "2026-10-01",
      "until": "2026-10-07"
    },
    "time_increment": 1
  }
}'

Tool catalog

97 tools. Pass a tool's name to POST /v1/tools/{name}, and an action from its list in arguments. Tools marked confirm change data or send messages and need confirm: true. Integration tools act on the accounts you've connected in Settings → Connections.

Workspace

`manage_tasks`	Manage the user's task boards (the Tasks page): list/create/update/delete boards, create, update, move, archive and delete tasks, add comments, and set dependencies between tasks. `list_boardscreate_boardupdate_boarddelete_boardboard_membersjoin_boardlist_tasksget_taskcreate_taskcreate_tasksupdate_taskmove_taskarchive_taskunarchive_taskdelete_taskadd_commentlist_commentsdelete_commentadd_dependencyremove_dependency`
`manage_teams`	Manage the user's teams: list/get/create/update/delete teams, list and remove members, invite people by email, list and revoke pending invitations, and turn the team invite code on or off. `listgetcreateupdatedeletejoinlist_membersremove_memberleaveinvitelist_invitationsrevoke_invitationregenerate_invite_codetransfer_ownership`
`manage_contexts`	Manage the user's contexts (reusable workspaces with a system prompt, model and settings) and each context's knowledge base. `listgetcreateupdatearchiveunarchivedeletelist_entriesadd_entryadd_entriesupdate_entrydelete_entry`
`manage_memories`	Read and manage what you remember about the user across conversations. `listsearchrememberupdateforget`
`manage_conversations`	Organize the user's chat conversations: list, rename, archive/unarchive, delete, create folders and move conversations into them, and create or revoke a public share link. `listget_messagesrenamearchiveunarchivedeleteshareunsharelist_folderscreate_folderdelete_foldermove_to_foldermove_to_project`
`manage_projects`	The user's AI Chat projects — named groups of conversations that share instructions (system prompt). `listgetcreateupdatearchiveunarchivedeleteadd_conversationremove_conversationlist_conversations`
`search_conversations`	Search across all past conversations for relevant context.
`search_knowledge_base`	Search the user's knowledge bases (the entries saved under their contexts, including chunks of long documents) and their saved memories.
`get_user_profile`	Read the current user's profile, preferences, role, subscription tier, and conversation history summary.
`modify_settings`	Adjust the AI's own settings mid-conversation: temperature, max_tokens, model, system prompt, or streaming preference.
`send_notification`	The user's notifications. `sendlistmark_readmark_all_read`
`check_credits`	The user's MultiverseAI credits. `balancetransactionsbuy_linkreferral_code`

Prompts & workflows

`prompt_runner`	The Prompt Runner page, from chat. `runhistoryget`
`manage_queue`	Manage the user's prompt queue (Queue / Advanced Queue pages). `enqueuelistgetrunretrycancelset_prioritydeleteclear_finishedstats`
`workflow_builder`	Create and edit workflows: whole graphs at once, individual nodes/edges, workflow metadata, layout, validation, duplication and deletion. `creategetreplace_graphadd_nodesadd_edgesupdate_noderemove_noderemove_edgeupdate_workflowauto_layoutvalidateduplicatedeletelist_folderscreate_folderdelete_foldermove_to_folderlist_sharesshareunsharelist_versionssave_versionrestore_version`
`execute_workflow`	Run a saved workflow end-to-end right now, as the user, and wait for the result (up to ~90s).
`execute_node`	Execute a single workflow node directly — without building a full workflow.
`list_workflows`	List the user's workflows with id, name, description, status, and node count.
`list_executions`	List recent workflow execution runs.
`get_execution_status`	Fetch the full state of a workflow execution by id — including per-node trace, outputs, and error details.
`list_node_types`	List every node type the agent can execute (via execute_node or as part of a workflow).
`describe_node_type`	Get the config schema for a node type: which keys belong on node.data, which are required, an example, and what the node outputs.
`schedule_action`	Schedule a saved workflow to run automatically: on a recurring cron schedule in a time zone, or once at a specific date/time. `createlistupdatepauseresumedelete`
`use_template`	Browse workflow templates (the user's own and the community gallery) and create a new workflow from one. `listsearchgetinstantiatepublish`
`list_integrations`	List what the user has connected: social/messaging accounts (Facebook Page, Instagram, WhatsApp, Threads, LinkedIn, X, TikTok, Telegram bots), Google, GitHub, their own AI provider keys, and configured integrations (SMTP, Slack, webhooks).

Agents

`manage_agent`	Create, inspect, update, archive or delete the user's custom AI agents, and attach tools to them. `creategetlistupdatearchivedeleteattach_tooldetach_toollist_toolslink_workflowunlink_workflowschedule_tasklist_tasksupdate_taskdelete_taskrunlist_versionsrestore_versionadd_test_caselist_test_casesdelete_test_caserun_testslist_sharesshareunshare`
`spawn_agent`	Spawn a sub-agent to handle a specific subtask.
`handoff_agent`	Hand off the current conversation to a specialized agent persona.

Social & messaging

`social_post_scheduler` confirm	Cross-platform social posting as the user's own connected accounts (X/Twitter, TikTok, Instagram, Facebook Page, Threads, LinkedIn). `publish_nowschedule_postlist_postsget_postcancel_postupdate_postlist_platforms`
`project_channels`	Read a project's connected channels and Google workspace (read-only). `list_channelsgmail_listgmail_getcalendar_listdrive_listdrive_get`
`project_channels_act` confirm	Act through a project's own connected channels (needs the user's approval). `send_messagegmail_sendcalendar_createcalendar_deleteconnect_link`
`twitter` confirm	Act as the user's own connected X (Twitter) account: post tweets (with images/video via media_urls) and threads, reply, quote, retweet, like, bookmark, delete; follow/unfollow/mute/block; read timeline, mentions, followers, bookmarks, likes `post_tweetpublish_threadreply_tweetquote_tweetretweetlike_tweetunlike_tweetbookmark_tweetdelete_tweetget_tweetget_analyticssearch_tweetsget_userget_timelineget_mentionsget_bookmarksget_liked_tweetsfollow_userunfollow_usermute_userblock_userget_followersget_followingsend_dmget_dmsupload_mediaget_listsget_list_tweetscreate_list`
`instagram` confirm	Act as the user's own connected Instagram Business/Creator account: publish photos, reels, carousels (2–10 items) and stories; list media; reply to or delete comments; delete media; media and account insights; hashtag search; posts the acco `get_profilelist_mediaget_media_insightsget_account_insightslist_commentsreply_commentdelete_commentcreate_postcreate_carouselcreate_storydelete_mediahashtag_searchget_mentionsbusiness_discoverylist_stories`
`facebook_page` confirm	Act on the user's own Facebook Pages: create text/link/photo/video posts, schedule posts and list scheduled ones, edit and delete posts, read posts, list and reply to comments, Page and per-post insights. `list_pagesget_pagecreate_postcreate_video_postschedule_postlist_scheduledlist_postsget_postupdate_postdelete_postlist_commentsreply_commentget_insightsget_post_insights`
`facebook_groups` confirm	Facebook Groups on the user's own connected Facebook account (Connections → Meta): list their groups, read a group's details, feed and members, and publish a post (text and/or link) to a group. `list_groupsget_groupget_feedget_memberspublish_post`
`threads` confirm	Post and read on the user's own connected Threads account (connect it in Connections → Social). `publish_postpublish_carouselreplylist_postsget_repliesget_insightsget_profiledelete_post`
`linkedin` confirm	Post on the user's own connected LinkedIn profile (connect it in Settings → Social). `publish_postcommentreactdelete_postget_profile`
`tiktok` confirm	Organic TikTok content on the user's own connected TikTok account (connect it in Connections → Social): publish a video or a photo post, check publish status, list and inspect videos, read and reply to comments, get hashtag suggestions. `publish_videopublish_photopublish_statuslist_videosget_videolist_commentsreply_commenthashtag_suggestions`
`messenger` confirm	Messenger (the user's Facebook Page inbox), Instagram DMs and WhatsApp on the user's own connected Meta account: read conversations and messages, look up a Messenger sender, and send text or media. `send_messagesend_mediasend_templatelist_conversationsget_messagesget_user_profile`
`whatsapp` confirm	WhatsApp Business on the user's own connected number (Connections → Meta). `send_messagesend_mediasend_templatelist_templatescreate_templatedelete_templateget_business_profilelist_numbersselect_number`
`telegram` confirm	The user's own Telegram bots, paired in Connections → Telegram: list the paired bots and send a message from one to a chat. `list_botssend_message`
`meeting_notetaker` confirm	AI meeting notetaker (Recall.ai): send a bot named "MultiverseAI Notetaker" into the user's Google Meet, Zoom, Microsoft Teams or Webex call (now, or scheduled with join_at), and read the resulting notes — summary, action items, decisions a `join_meetinglist_meetingsget_meetingleave_meeting`
`youtube`	Query the YouTube Data API: search videos/channels, fetch video metadata and stats, list a channel's videos, get playlist contents, and read top comments. `search_videossearch_channelsget_videoget_channellist_channel_videosget_playlistget_comments`
`youtube_transcript`	Read what's inside a YouTube video via QuickTube: the full transcript, an AI summary, the best sections (chapter-style highlights with timestamps), the comment threads, or a download link. `transcriptsummarizebest_sectionscommentsdownload_url`
`producthunt`	Query Product Hunt for trending products, search by topic, fetch a specific product by slug. `trendingsearchget_productby_topic`

Ads & commerce

`facebook_ads` confirm	Run Meta (Facebook/Instagram) ads as an ads manager through the Marketing API. `list_ad_accountslist_pageslist_pixelslist_campaignsget_campaignlist_adsetslist_adslist_creativeslist_audiencesget_insightscreate_campaigncreate_adsetupload_imagecreate_ad_creativecreate_adcreate_audienceupdatepauseresumedelete`
`facebook_fundraiser` confirm	Create and manage Facebook Page fundraisers for charitable organizations. `list_fundraisersget_fundraisercreate_fundraiserend_fundraiser`
`meta_catalog` confirm	Meta Commerce Manager catalogs (Facebook/Instagram Shops) on the user's own connected Meta account: list or create catalogs for a business, list/add/update/delete products, list or create product sets. `list_catalogscreate_cataloglist_productsadd_productupdate_productdelete_productlist_product_setscreate_product_set`
`tiktok_ads` confirm	Manage TikTok advertising campaigns. `list_campaignsget_campaigncreate_campaignupdate_campaignlist_adgroupscreate_adgroupupdate_adgrouplist_adscreate_adupdate_adpauseresume`
`tiktok_analytics`	Pull TikTok advertising performance data including campaign metrics, ROAS, spend, impressions, clicks, and conversions. `campaign_reportadgroup_reportad_reportaudience_insightsrealtime_stats`
`google_ads` confirm	Manage Google Ads campaigns. `list_campaignsget_campaigncreate_campaignupdate_campaignkeyword_researchget_performancelist_adgroupsadjust_bidpauseresume`
`google_analytics`	Pull Google Analytics data including traffic, conversions, audience demographics, page views, and custom reports. `traffic_overviewpage_viewsconversionsaudiencerealtimecustom_report`
`shopify`	Query Shopify store data. `list_productsget_productlist_ordersget_orderlist_customersget_inventorysearch_productsget_analytics`
`stripe` confirm	Not available yet: connecting your own Stripe account is not supported, so every action returns an error. `list_chargesget_chargelist_invoicescreate_invoiceget_balancelist_subscriptionslist_customersget_customerlist_payoutsrefund`
`crm` confirm	Read and write CRM data (HubSpot or Salesforce). `list_contactsget_contactcreate_contactupdate_contactlist_dealsget_dealcreate_dealupdate_deallist_pipelinessearchlist_activities`

Google, email & productivity

`email` confirm	Send email or save a draft from the user's own connected Gmail (Connections → Google Drive). `draftsendread_inboxsearchget_thread`
`send_email` confirm	Send a transactional email via Resend (different from gmail tool).
`resend` confirm	Resend email platform using the user's connected Resend account (Integrations → Resend). `sendsend_batchlist_emailsget_emailcancel_emaillist_domainsverify_domaincreate_contactlist_contactslist_segments`
`angelsend`	Read the user's AngelSend email-outreach account (read-only, no approval): campaigns and their stats, the send queue (per-row BCC, compliance-archive status, bounce and unsubscribe details), recipients with pause/stop state, merge variables `search_actionsdescribe_actioncallconnections`
`angelsend_act` confirm	Change things in the user's AngelSend account (needs the user's approval). `callconnectdisconnect`
`google_docs` confirm	Create Google Docs (optionally with content, in a folder) and add text to them, in the user's own Google account. `createreadupdatesearchappendreplace_text`
`google_sheets`	Read and write Google Sheets data. `readwriteappendcreatelist_sheetsclearbatch_update`
`google_drive` confirm	Create, upload, update, copy, move, share and delete files and folders in the user's own Google Drive. `createuploadupdatedeletecopymovesharecreate_foldersearchlistgetget_contentlist_shared`
`google_photos`	Search and browse Google Photos. `list_albumsget_albumlist_mediaget_mediasearch_media`
`calendar` confirm	Manage calendar events. `list_eventscreate_eventupdate_eventdelete_eventcheck_availabilityfind_free_time`
`slack` confirm	Interact with Slack. `send_messageread_channellist_channelssearch_messagesget_threadreactupload_file`
`discord` confirm	Send and read Discord messages. `send_messagesend_via_webhooklist_guildslist_channelsread_channeldelete_message`
`github` confirm	Work in the user's GitHub repos through their MultiverseAI GitHub App installation: list repos, issues and PRs, read files, create issues, comments, branches, files and pull requests. `list_installationslist_reposlist_issuescreate_issueadd_commentlist_prscreate_prget_fileget_contentscreate_branchwrite_fileimport_repo_contextget_repoget_issuesearch_codelist_commitsget_actions_status`
`jira` confirm	Manage Jira issues. `list_projectssearch_issuesget_issuecreate_issueupdate_issueadd_commentlist_transitionstransition`
`linear` confirm	Interact with Linear. `list_issuessearch_issuesget_issuecreate_issueupdate_issueadd_commentlist_teamslist_projectslist_states`
`notion` confirm	Interact with Notion. `searchget_pageget_page_contentcreate_pageupdate_pageappend_contentquery_databaseget_database`
`webhook` confirm	Trigger arbitrary webhooks to connect with external apps (Zapier, Make, n8n).

Web & research

`web_search`	Search the web in real-time and return results with source citations.
`web_scrape`	Fetch and extract content from a URL.
`web_capture`	Render a public web page in a real browser and save it as a PNG screenshot or a PDF; returns a hosted URL (plus markdown to show a screenshot inline). `screenshotpdf`
`firecrawl`	Advanced web scraping with Firecrawl. `scrapecrawlmapsearch`
`perplexity_search`	AI-powered web search via Perplexity.
`semantic_search`	Advanced semantic / vector search across your indexed knowledge.
`vector_store` confirm	Write to and search the user's OWN vector database (Pinecone or Qdrant): index text documents (embedded with the user's own embedding key, OpenAI text-embedding-3-small by default), query by meaning, or delete records by id or metadata filt `indexquerydelete`

Media, code & data

`generate_image`	Generate an image from a text description, or list the image models (action "list_models"). `generatelist_models`
`generate_video`	Start generating a short video clip (text-to-video, or image-to-video from a start image URL). `generatestatuslistcancel`
`heygen`	HeyGen AI avatar videos. `list_avatarslist_voiceslist_languagescreate_videotranslate_videoget_video`
`tavus`	Tavus AI replicas: personalized talking-head videos and live AI video calls. `list_replicaslist_personascreate_videoget_videocreate_conversationget_conversationend_conversation`
`text_to_speech`	Turn text into natural speech (ElevenLabs) and return a hosted mp3 URL plus a markdown player link to include in your reply.
`elevenlabs`	ElevenLabs audio. `list_voicestext_to_speechspeech_to_textsound_effect`
`multi_model`	Ask the same prompt to several AI providers (OpenAI, Anthropic, Google) and get every answer side by side — to compare, cross-check facts by consensus, pick the best of n, or fall back when one provider fails.
`generate_chart`	Generate a chart from data.
`execute_code` confirm	Execute a JavaScript code snippet in an isolated sandbox (no network, files or secrets).
`code_execute` confirm	Execute JavaScript code in a sandboxed environment and return the result.
`code_analyze`	Analyze code for potential issues, complexity, and provide suggestions.
`calculator`	Evaluate a mathematical expression.
`datetime`	Get the current date and time, or convert between timezones.
`json_transform`	Parse, format, filter, or transform JSON data.
`summarize`	Generate a concise summary of provided text.
`parse_file`	Parse and extract content from uploaded files (CSV, PDF, TXT, JSON, images).
`database_query`	Read rows from the user's MultiverseAI data (read-only).
`query_analytics`	Pull analytics data: ad performance, token usage, cost breakdowns, or conversation statistics.

Link Previews

Title, description and image metadata for URLs.

POST

/v1/link-previews

Fetch link previews

Each preview has url and, when the page declares them, title, description, image and siteName. URLs that cannot be previewed map to null.

Request body

Field	Type	Required	Description
`urls`	`string[]`	Yes	URLs to preview

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/link-previews" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "urls": [
    "https://example.com"
  ]
}'

Credits

Your MultiverseAI credit balance. AI calls made with a key are paid from its owner's credits; when the balance runs out they fail with 402 INSUFFICIENT_CREDITS.

GET

/v1/credits

Get your credit balance

Query parameters

Field	Type	Required	Description
`fresh`	`boolean`	No	true skips the 60-second balance cacheDefault: `false`

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/credits?fresh=true" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

GET

/v1/credits/transactions

List your credit transactions, newest first

Returns 502 UPSTREAM_ERROR when the credit ledger is temporarily unreachable.

Query parameters

Field	Type	Required	Description
`limit`	`integer`	No	How many, up to 100Default: `50`

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/credits/transactions?limit=2" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

Workflows

Your saved workflows. Runs use the same server-side engine as the app, as the key's owner.

GET

/v1/workflows

List workflows

Key permission: workflows

Query parameters

Field	Type	Required	Description
`status`	`string`	No	Filter by workflow status
`limit`	`integer`	No	Page size, up to 100Default: `50`
`offset`	`integer`	No	Items to skipDefault: `0`

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/workflows?limit=10" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

POST

/v1/workflows

Create a workflow

Key permission: workflows

nodes and edges use the same JSON as the builder's export (see the Workflow JSON docs). Node ids must be unique and every edge must join two of them.
New workflows are drafts unless is_active is true. The graph may be up to 2 MB and 1000 nodes.

Request body

Field	Type	Required	Description
`name`	`string`	Yes	Name
`description`	`string`	No	Description
`nodes`	`object[]`	No	Builder nodes: { id, type, data?, position? } (position defaults to a left-to-right layout)
`edges`	`object[]`	No	Builder edges: { source, target, sourceHandle?, id? } between node ids
`category`	`string`	No	Category
`tags`	`string[]`	No	Tags
`folder_id`	`uuid`	No	Folder
`is_active`	`boolean`	No	Activate (also sets status to "active" / "draft"), like the builder's Active switch

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/workflows" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "name": "Daily digest",
  "nodes": [
    {
      "id": "trigger",
      "type": "trigger",
      "data": {
        "triggerType": "manual"
      }
    },
    {
      "id": "summarise",
      "type": "aiPrompt",
      "data": {
        "prompt": "Summarise {{trigger.topic}} in 5 bullets"
      }
    }
  ],
  "edges": [
    {
      "source": "trigger",
      "target": "summarise"
    }
  ]
}'

GET

/v1/workflows/{id}

Get a workflow, including nodes and edges

Key permission: workflows

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/workflows/3f8a1c2d-5e6b-4a7c-8d9e-0f1a2b3c4d5e" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

PATCH

/v1/workflows/{id}

Update a workflow

Key permission: workflows

Sending nodes or edges replaces that part of the graph (edges are checked against the saved nodes when only edges are sent) and bumps version.

Request body

Field	Type	Required	Description
`name`	`string`	No	Name
`description`	`string`	No	Description
`nodes`	`object[]`	No	Builder nodes: { id, type, data?, position? } (position defaults to a left-to-right layout)
`edges`	`object[]`	No	Builder edges: { source, target, sourceHandle?, id? } between node ids
`category`	`string`	No	Category
`tags`	`string[]`	No	Tags
`folder_id`	`uuid`	No	Folder
`is_active`	`boolean`	No	Activate (also sets status to "active" / "draft"), like the builder's Active switch

curl -X PATCH "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/workflows/3f8a1c2d-5e6b-4a7c-8d9e-0f1a2b3c4d5e" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "is_active": true
}'

DELETE

/v1/workflows/{id}

Delete a workflow

Key permission: workflows

curl -X DELETE "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/workflows/3f8a1c2d-5e6b-4a7c-8d9e-0f1a2b3c4d5e" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

POST

/v1/workflows/{id}/run

Run a workflow

Key permission: workflows

By default the run starts in the background and the call returns 202 with an execution id; poll status_url (GET /v1/executions/{id}).
POST /v1/workflows/{id}/execute is an alias.

Request body

Field	Type	Required	Description
`inputs`	`object`	No	Values passed to the workflow triggerDefault: `{}`
`wait`	`boolean`	No	true waits for the run to finish and returns its resultDefault: `false`

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/workflows/3f8a1c2d-5e6b-4a7c-8d9e-0f1a2b3c4d5e/run" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "inputs": {
    "topic": "Q3 sales"
  }
}'

GET

/v1/workflows/{id}/executions

List a workflow's runs

Key permission: workflows

Query parameters

Field	Type	Required	Description
`limit`	`integer`	No	Page size, up to 100Default: `50`
`offset`	`integer`	No	Items to skipDefault: `0`

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/workflows/3f8a1c2d-5e6b-4a7c-8d9e-0f1a2b3c4d5e/executions" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

Executions

Workflow runs across all of your workflows.

GET

/v1/executions

List runs

Key permission: workflows

Query parameters

Field	Type	Required	Description
`workflow_id`	`uuid`	No	Only runs of this workflow
`status`	`string`	No	Filter by run status
`limit`	`integer`	No	Page size, up to 100Default: `50`
`offset`	`integer`	No	Items to skipDefault: `0`

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/executions?workflow_id=3f8a1c2d-5e6b-4a7c-8d9e-0f1a2b3c4d5e" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

GET

/v1/executions/{id}

Get a run, including execution_data, node_outputs and error_message

Key permission: workflows

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/executions/b7c8d9e0-1f2a-4b3c-9d4e-5f6a7b8c9d0e" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

POST

/v1/executions/{id}/cancel

Cancel a running execution

Key permission: workflows

The run stops before its next node; a node already in flight (e.g. an AI call) finishes first. Its status becomes "cancelled".
A run that already finished returns 409 INVALID_STATE. Also available as POST /v1/workflows/{id}/executions/{execution_id}/cancel.

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/executions/b7c8d9e0-1f2a-4b3c-9d4e-5f6a7b8c9d0e/cancel" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

Scheduling Runs

Workflows with a schedule trigger are run automatically by MultiverseAI's internal scheduler; there is nothing to call. The scheduler function itself only accepts the platform's internal cron secret and cannot be called with an API key.

To run a workflow on demand, use POST /v1/workflows/{id}/run.
To run a workflow (or a prompt) at a specific time, add it to the queue with POST /v1/queue and a future scheduled_for — it waits as "scheduled" and runs then.

POST

/v1/queue

Schedule a workflow run for later

Key permission: workflows

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/queue" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "workflow_id": "3f8a1c2d-5e6b-4a7c-8d9e-0f1a2b3c4d5e",
  "scheduled_for": "2026-10-08T09:00:00Z"
}'

Agents

Your AI agents. Chat uses the agent's own instructions, model, knowledge context and workflow triggers, and every exchange is saved to the agent's history under a conversation_id.

GET

/v1/agents

List agents

Key permission: agents

Query parameters

Field	Type	Required	Description
`include_archived`	`boolean`	No	Include archived agentsDefault: `false`
`limit`	`integer`	No	Page size, up to 100Default: `50`
`offset`	`integer`	No	Items to skipDefault: `0`

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/agents" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

POST

/v1/agents

Create an agent

Key permission: agents

Starts as a draft with response_length "balanced" and the app's default model unless you set them.

Request body

Field	Type	Required	Description
`name`	`string`	Yes	Name
`description`	`string`	No	Description
`instructions`	`string`	No	System instructions
`default_model`	`string`	No	Model id (GET /v1/models)
`context_id`	`uuid`	No	Knowledge context (GET /v1/contexts)
`personality_template`	`string`	No	Personality template name
`response_length`	`string`	No	"concise", "balanced" or "detailed"
`max_tokens_limit`	`integer`	No	Token cap per reply
`max_cost_limit`	`number`	No	Cost cap
`tags`	`string[]`	No	Tags
`image_url`	`string`	No	https:// avatar URL
`voice_id`	`string`	No	Voice for text-to-speech
`voice_settings`	`object`	No	Voice settings
`status`	`string`	No	"draft", "published" or "archived"

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/agents" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "name": "Support Assistant",
  "instructions": "Answer product questions briefly and link the docs.",
  "response_length": "concise",
  "tags": [
    "support"
  ]
}'

GET

/v1/agents/{id}

Get an agent, including its instructions and configuration

Key permission: agents

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/agents/5a6b7c8d-9e0f-4a1b-8c2d-3e4f5a6b7c8d" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

PATCH

/v1/agents/{id}

Update an agent

Key permission: agents

Only the owner can change an agent; agents shared with you can be read and chatted with, and return 404 here.

Request body

Field	Type	Required	Description
`name`	`string`	No	Name
`description`	`string`	No	Description
`instructions`	`string`	No	System instructions
`default_model`	`string`	No	Model id (GET /v1/models)
`context_id`	`uuid`	No	Knowledge context (GET /v1/contexts)
`personality_template`	`string`	No	Personality template name
`response_length`	`string`	No	"concise", "balanced" or "detailed"
`max_tokens_limit`	`integer`	No	Token cap per reply
`max_cost_limit`	`number`	No	Cost cap
`tags`	`string[]`	No	Tags
`image_url`	`string`	No	https:// avatar URL
`voice_id`	`string`	No	Voice for text-to-speech
`voice_settings`	`object`	No	Voice settings
`status`	`string`	No	"draft", "published" or "archived"
`is_archived`	`boolean`	No	Archive or unarchive

curl -X PATCH "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/agents/5a6b7c8d-9e0f-4a1b-8c2d-3e4f5a6b7c8d" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "status": "published"
}'

DELETE

/v1/agents/{id}

Delete an agent and its chat history

Key permission: agents

curl -X DELETE "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/agents/5a6b7c8d-9e0f-4a1b-8c2d-3e4f5a6b7c8d" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

POST

/v1/agents/{id}/chat

Chat with an agent

Key permission: agents

Request body

Field	Type	Required	Description
`message`	`string`	Yes	Your message
`conversation_id`	`uuid`	No	Continue a saved conversation (its last 40 turns are used as context). Omit to start a new one; the response returns its id.
`conversation_history`	`array`	No	Explicit [{ role: "user" \| "assistant", content }] history; used instead of the saved history
`stream`	`boolean`	No	Stream chat.completion.chunk events (each carries conversation_id), ending with [DONE]Default: `false`

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/agents/5a6b7c8d-9e0f-4a1b-8c2d-3e4f5a6b7c8d/chat" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "message": "How do I reset my password?"
}'

GET

/v1/agents/{id}/chat

Read an agent's saved messages

Key permission: agents

Query parameters

Field	Type	Required	Description
`conversation_id`	`uuid`	No	Only this conversation
`limit`	`integer`	No	Page size, up to 200Default: `50`
`offset`	`integer`	No	Items to skipDefault: `0`

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/agents/5a6b7c8d-9e0f-4a1b-8c2d-3e4f5a6b7c8d/chat?conversation_id=c1d2e3f4-a5b6-4c7d-8e9f-0a1b2c3d4e5f" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

Conversations

Your chat history — the same conversations the MultiverseAI apps show. Sending a message works like the app: your message is saved, the model answers with the conversation (last 40 turns) as context, and the reply is saved too.

GET

/v1/conversations

List conversations

Query parameters

Field	Type	Required	Description
`archived`	`boolean`	No	true lists archived conversations insteadDefault: `false`
`q`	`string`	No	Search titles
`limit`	`integer`	No	Page size, up to 100Default: `50`
`offset`	`integer`	No	Items to skipDefault: `0`

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/conversations?q=launch" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

POST

/v1/conversations

Create a conversation

Returns 201 with the conversation.

Request body

Field	Type	Required	Description
`title`	`string`	No	TitleDefault: `New Conversation`
`model`	`string`	No	Default model for the conversationDefault: `openrouter/openai/gpt-6-luna`
`folder_id`	`uuid`	No	Folder to put it in

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/conversations" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "title": "Launch plan",
  "model": "openrouter/anthropic/claude-opus-5.5"
}'

GET

/v1/conversations/{id}

Get a conversation

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/conversations/2e3f4a5b-6c7d-4e8f-9a0b-1c2d3e4f5a6b" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

PATCH

/v1/conversations/{id}

Update a conversation

Request body

Field	Type	Required	Description
`title`	`string`	No	Title
`model`	`string`	No	Default model
`folder_id`	`uuid \| null`	No	Folder (null removes it)
`is_archived`	`boolean`	No	Archive or unarchive

curl -X PATCH "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/conversations/2e3f4a5b-6c7d-4e8f-9a0b-1c2d3e4f5a6b" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "is_archived": true
}'

DELETE

/v1/conversations/{id}

Delete a conversation

curl -X DELETE "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/conversations/2e3f4a5b-6c7d-4e8f-9a0b-1c2d3e4f5a6b" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

GET

/v1/conversations/{id}/messages

List messages (oldest first)

Query parameters

Field	Type	Required	Description
`limit`	`integer`	No	Page size, up to 200Default: `100`
`offset`	`integer`	No	Items to skipDefault: `0`

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/conversations/2e3f4a5b-6c7d-4e8f-9a0b-1c2d3e4f5a6b/messages" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

POST

/v1/conversations/{id}/messages

Send a message and get the reply

Returns 201 { object: "conversation.exchange", conversation_id, user_message, message }.
Streaming sends { object: "conversation.message" } (your saved message), then { object: "conversation.delta", delta } chunks and { object: "conversation.tool_event", event, tool_call } events, then the saved reply as another conversation.message, then data: [DONE].

Request body

Field	Type	Required	Description
`content`	`string`	Yes	Your message
`model`	`string`	No	Model for this reply (defaults to the conversation's model)
`system_prompt`	`string`	No	System prompt for this reply
`stream`	`boolean`	No	Stream the exchange over SSEDefault: `false`
`temperature`	`number`	No	Sampling temperature, 0–2
`max_tokens`	`integer`	No	Maximum tokens to generate (max_completion_tokens is accepted too)
`top_p`	`number`	No	Nucleus sampling, 0–1
`frequency_penalty`	`number`	No	-2 to 2
`presence_penalty`	`number`	No	-2 to 2
`tools`	`true \| string[]`	No	true enables every MultiverseAI agent tool; an array enables only the named tools (names from GET /v1/tools). Tools run server-side as you. OpenAI-style function definitions are rejected with UNSUPPORTED_TOOLS.
`max_tool_iterations`	`integer`	No	Cap on tool-call rounds when tools are enabled (1–20)

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/conversations/2e3f4a5b-6c7d-4e8f-9a0b-1c2d3e4f5a6b/messages" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "content": "Draft three launch-day tweets.",
  "tools": true
}'

Contexts

Prompt contexts: reusable system prompts with model and sampling settings.

GET

/v1/contexts

List contexts

Query parameters

Field	Type	Required	Description
`include_archived`	`boolean`	No	Include archived contextsDefault: `false`
`limit`	`integer`	No	Page size, up to 100Default: `50`
`offset`	`integer`	No	Items to skipDefault: `0`

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/contexts" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

POST

/v1/contexts

Create a context

Returns 201 with the context.

Request body

Field	Type	Required	Description
`name`	`string`	Yes	Name
`description`	`string`	No	Description
`color`	`string`	No	Hex colorDefault: `#6366f1`
`tags`	`string[]`	No	Tags
`model`	`string`	No	Model id
`temperature`	`number`	No	Sampling temperature
`max_tokens`	`integer`	No	Max tokens
`system_prompt`	`string`	No	System prompt
`metadata`	`object`	No	Free-form JSON

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/contexts" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "name": "Brand voice",
  "system_prompt": "Write in a friendly, plain-spoken tone.",
  "model": "openrouter/openai/gpt-6-luna",
  "temperature": 0.7,
  "tags": [
    "marketing"
  ]
}'

GET

/v1/contexts/{id}

Get a context

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/contexts/7d8e9f0a-1b2c-4d3e-8f4a-5b6c7d8e9f0a" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

PATCH

/v1/contexts/{id}

Update a context

Request body

Field	Type	Required	Description
`name`	`string`	No	Name (required on create)
`description`	`string`	No	Description
`color`	`string`	No	Hex colorDefault: `#6366f1`
`tags`	`string[]`	No	Tags
`model`	`string`	No	Model id
`temperature`	`number`	No	Sampling temperature
`max_tokens`	`integer`	No	Max tokens
`system_prompt`	`string`	No	System prompt
`metadata`	`object`	No	Free-form JSON
`is_archived`	`boolean`	No	Archive or unarchive

curl -X PATCH "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/contexts/7d8e9f0a-1b2c-4d3e-8f4a-5b6c7d8e9f0a" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "temperature": 0.4
}'

DELETE

/v1/contexts/{id}

Delete a context

curl -X DELETE "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/contexts/7d8e9f0a-1b2c-4d3e-8f4a-5b6c7d8e9f0a" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

Boards

Task boards you own or share through a team.

GET

/v1/boards

List boards

Key permission: tasks

Query parameters

Field	Type	Required	Description
`limit`	`integer`	No	Page size, up to 100Default: `50`
`offset`	`integer`	No	Items to skipDefault: `0`

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/boards" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

POST

/v1/boards

Create a board

Key permission: tasks

Request body

Field	Type	Required	Description
`name`	`string`	Yes	Name
`description`	`string`	No	Description
`color`	`string`	No	Color
`team_id`	`uuid`	No	Create it as a team board (create only)

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/boards" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "name": "Marketing"
}'

GET

/v1/boards/{id}

Get a board

Key permission: tasks

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/boards/4c5d6e7f-8a9b-4c0d-9e1f-2a3b4c5d6e7f" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

PATCH

/v1/boards/{id}

Update a board

Key permission: tasks

Request body

Field	Type	Required	Description
`name`	`string`	No	Name
`description`	`string`	No	Description
`color`	`string`	No	Color

curl -X PATCH "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/boards/4c5d6e7f-8a9b-4c0d-9e1f-2a3b4c5d6e7f" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "name": "Product roadmap"
}'

DELETE

/v1/boards/{id}

Delete a board

Key permission: tasks

curl -X DELETE "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/boards/4c5d6e7f-8a9b-4c0d-9e1f-2a3b4c5d6e7f" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

GET

/v1/boards/{id}/tasks

List a board's tasks (by position, archived tasks excluded)

Key permission: tasks

Query parameters

Field	Type	Required	Description
`limit`	`integer`	No	Page size, up to 200Default: `100`
`offset`	`integer`	No	Items to skipDefault: `0`

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/boards/4c5d6e7f-8a9b-4c0d-9e1f-2a3b4c5d6e7f/tasks" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

Tasks

Tasks on your boards.

GET

/v1/tasks

List tasks

Key permission: tasks

Query parameters

Field	Type	Required	Description
`board_id`	`uuid`	No	Only this board
`status`	`string`	No	Filter by status
`priority`	`string`	No	Filter by priority
`assignee`	`string`	No	Filter by assignee
`include_archived`	`boolean`	No	Include archived tasksDefault: `false`
`limit`	`integer`	No	Page size, up to 100Default: `50`
`offset`	`integer`	No	Items to skipDefault: `0`

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/tasks?status=todo&priority=high" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

POST

/v1/tasks

Create a task

Key permission: tasks

Returns 201 with the task. Invalid enum values return 400 INVALID_VALUE.

Request body

Field	Type	Required	Description
`title`	`string`	Yes	Title
`description`	`string`	No	Description
`status`	`string`	No	backlog \| todo \| in-progress \| review \| doneDefault: `todo`
`priority`	`string`	No	low \| medium \| high \| urgentDefault: `medium`
`assignee`	`string`	No	Assignee
`tags`	`string[]`	No	Tags
`due_date`	`string`	No	ISO 8601 date
`board_id`	`uuid`	No	Board (your default board if omitted on create)
`estimated_hours`	`number`	No	Estimate in hours

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/tasks" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "title": "Write launch blog post",
  "priority": "high",
  "due_date": "2026-10-15",
  "tags": [
    "launch"
  ]
}'

GET

/v1/tasks/{id}

Get a task

Key permission: tasks

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/tasks/8e9f0a1b-2c3d-4e4f-a5b6-c7d8e9f0a1b2" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

PATCH

/v1/tasks/{id}

Update a task

Key permission: tasks

Request body

Field	Type	Required	Description
`title`	`string`	No	Title (required on create)
`description`	`string`	No	Description
`status`	`string`	No	backlog \| todo \| in-progress \| review \| doneDefault: `todo`
`priority`	`string`	No	low \| medium \| high \| urgentDefault: `medium`
`assignee`	`string`	No	Assignee
`tags`	`string[]`	No	Tags
`due_date`	`string`	No	ISO 8601 date
`board_id`	`uuid`	No	Board (your default board if omitted on create)
`estimated_hours`	`number`	No	Estimate in hours
`logged_hours`	`number`	No	Hours logged
`position`	`number`	No	Order on the board

curl -X PATCH "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/tasks/8e9f0a1b-2c3d-4e4f-a5b6-c7d8e9f0a1b2" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "status": "in-progress",
  "logged_hours": 1.5
}'

DELETE

/v1/tasks/{id}

Delete a task

Key permission: tasks

curl -X DELETE "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/tasks/8e9f0a1b-2c3d-4e4f-a5b6-c7d8e9f0a1b2" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

Teams

Teams you belong to (read-only).

GET

/v1/teams

List your teams

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/teams" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

GET

/v1/teams/{id}

Get a team with its members

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/teams/0a1b2c3d-4e5f-4a6b-8c7d-9e0f1a2b3c4d" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

Prompt Queue

The prompt queue: run a prompt (or a saved workflow) in the background now, or schedule it for later. Items and their responses appear in the app's queue history.

POST

/v1/queue

Add an item to the queue

Key permission: workflows

Returns 201 with the item. Unscheduled items start right away (status "running") unless run is false.

Request body

Field	Type	Required	Description
`prompt`	`string`	No	The prompt (required unless workflow_id is given)
`workflow_id`	`uuid`	No	Run this saved workflow instead of a prompt
`model`	`string`	No	Model idDefault: `openrouter/openai/gpt-6-luna`
`priority`	`string`	No	critical \| high \| normal \| lowDefault: `normal`
`destination`	`string`	No	Where the result should go
`scheduled_for`	`string`	No	ISO 8601 time in the future; the item waits as "scheduled" and runs then
`run`	`boolean`	No	false only enqueues it (status pending); start it later with POST /v1/queue/{id}/runDefault: `true`
`wait`	`boolean`	No	Run now and wait for the resultDefault: `false`

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/queue" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "prompt": "Summarise this week in AI news.",
  "priority": "high",
  "scheduled_for": "2026-10-08T09:00:00Z"
}'

GET

/v1/queue

List queue items

Key permission: workflows

Query parameters

Field	Type	Required	Description
`status`	`string`	No	Comma-separated statuses, e.g. pending,running,scheduled
`limit`	`integer`	No	Page size, up to 100Default: `50`
`offset`	`integer`	No	Items to skipDefault: `0`

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/queue?status=pending,running,scheduled" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

GET

/v1/queue/{id}

Get a queue item (including its response once done)

Key permission: workflows

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/queue/6f7a8b9c-0d1e-4f2a-b3c4-d5e6f7a8b9c0" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

POST

/v1/queue/{id}/run

Run a waiting item now

Key permission: workflows

Request body

Field	Type	Required	Description
`wait`	`boolean`	No	Wait for the resultDefault: `false`

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/queue/6f7a8b9c-0d1e-4f2a-b3c4-d5e6f7a8b9c0/run" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "wait": true
}'

DELETE

/v1/queue/{id}

Cancel a waiting item

Key permission: workflows

Only pending, scheduled or retrying items can be cancelled (otherwise 400 INVALID_STATE). Returns the item with status "cancelled".

curl -X DELETE "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/queue/6f7a8b9c-0d1e-4f2a-b3c4-d5e6f7a8b9c0" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

Social Posts

Publish or schedule posts as your own connected accounts (connect them in Settings first), or as a project's accounts with project_id / X-Project-Id (see Projects). Calling these endpoints is the go-ahead to post.

GET

/v1/platforms

Which posting platforms are connected

Key permission: posts

An Instagram, Threads or LinkedIn account whose access has lapsed shows connected: false with status: "expired" — reconnect Instagram and Threads in Settings → Meta, LinkedIn in Settings → Social (LinkedIn access lasts 60 days and cannot be refreshed).

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/platforms" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

POST

/v1/posts/create

Publish now

Key permission: posts

status is "published" when every platform succeeded, "partial" when some failed, "processing" (HTTP 202) when nothing failed but a platform is still processing the media, and "failed" (HTTP 502) when none succeeded.
Instagram and TikTok process video before it can go live. The call waits up to about a minute; after that the platform's result is "processing" (with its container_id or publish_id) and MultiverseAI publishes it as soon as processing finishes — follow it with GET /v1/posts/{id}/status or the post.published / post.failed webhooks. Instagram media is never published before Instagram reports it FINISHED.
Every create is recorded: the response id works with GET /v1/posts/{id}/status. Created posts don't appear in GET /v1/posts/scheduled.

Request body

Field	Type	Required	Description
`platforms`	`string[]`	Yes	Any of "twitter", "tiktok", "instagram", "facebook", "threads", "linkedin", "youtube"
`content`	`string`	No	Post text (content or media_urls is required)
`media_urls`	`string[]`	No	https:// media URLs. Instagram and TikTok need media; X (Twitter) takes up to 4 images or 1 video; Threads takes 1, or 2–10 as a carousel (text is required unless it is a carousel); LinkedIn takes 1–20 images or one video (text up to 3,000 characters), or a link via platform_options.linkedin.link instead of media; YouTube takes exactly one video.
`platform_options`	`object`	No	Per-platform options, e.g. { facebook: { page_id, link }, tiktok: { privacy_level, disable_comment }, threads: { account, alt_text }, linkedin: { link, link_title, visibility } }. linkedin.link shares a URL as an article post (no media alongside it), link_title overrides its title, and visibility is "PUBLIC" (default) or "CONNECTIONS". youtube: { title, privacy_status, tags, category_id, made_for_kids, publish_at, notify_subscribers, thumbnail_url } — title defaults to the first line of the text (100 characters max) and the text becomes the description; privacy_status is "private" (default), "unlisted" or "public" — until MultiverseAI's Google app passes YouTube's audit, uploads stay private whatever is asked; publish_at uploads it private and publishes it then. tiktok.privacy_level must be one the creator's account allows (default "SELF_ONLY"; unaudited TikTok apps can only post SELF_ONLY). <platform>.account (or account_id) picks one of several connected accounts by username or id.
`project_id`	`string`	No	Post as this project's accounts (see Projects). Also accepted as an X-Project-Id header.

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/posts/create" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "platforms": [
    "twitter",
    "facebook"
  ],
  "content": "We just shipped a new feature!"
}'

POST

/v1/posts/schedule

Schedule a post

Key permission: posts

Returns 201 { object: "scheduled_post", … }. Scheduled posts are published automatically at scheduled_at (within a minute); a platform still processing media leaves the post "processing" until it is live. Each platform's outcome also arrives as a post.published / post.failed webhook.

Request body

Field	Type	Required	Description
`platforms`	`string[]`	Yes	Any of "twitter", "tiktok", "instagram", "facebook", "threads", "linkedin", "youtube"
`content`	`string`	No	Post text (content or media_urls is required)
`media_urls`	`string[]`	No	https:// media URLs. Instagram and TikTok need media; X (Twitter) takes up to 4 images or 1 video; Threads takes 1, or 2–10 as a carousel (text is required unless it is a carousel); LinkedIn takes 1–20 images or one video (text up to 3,000 characters), or a link via platform_options.linkedin.link instead of media; YouTube takes exactly one video.
`platform_options`	`object`	No	Per-platform options, e.g. { facebook: { page_id, link }, tiktok: { privacy_level, disable_comment }, threads: { account, alt_text }, linkedin: { link, link_title, visibility } }. linkedin.link shares a URL as an article post (no media alongside it), link_title overrides its title, and visibility is "PUBLIC" (default) or "CONNECTIONS". youtube: { title, privacy_status, tags, category_id, made_for_kids, publish_at, notify_subscribers, thumbnail_url } — title defaults to the first line of the text (100 characters max) and the text becomes the description; privacy_status is "private" (default), "unlisted" or "public" — until MultiverseAI's Google app passes YouTube's audit, uploads stay private whatever is asked; publish_at uploads it private and publishes it then. tiktok.privacy_level must be one the creator's account allows (default "SELF_ONLY"; unaudited TikTok apps can only post SELF_ONLY). <platform>.account (or account_id) picks one of several connected accounts by username or id.
`project_id`	`string`	No	Post as this project's accounts (see Projects). Also accepted as an X-Project-Id header.
`scheduled_at`	`string`	Yes	ISO 8601 time in the future

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/posts/schedule" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "platforms": [
    "instagram"
  ],
  "content": "See you at the meetup tonight",
  "media_urls": [
    "https://example.com/banner.jpg"
  ],
  "scheduled_at": "2026-10-10T17:00:00Z"
}'

GET

/v1/posts/scheduled

List scheduled posts

Key permission: posts

Query parameters

Field	Type	Required	Description
`status`	`string`	No	pending \| publishing \| processing \| published \| failed \| cancelled
`limit`	`integer`	No	Up to 100Default: `50`

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/posts/scheduled?status=pending" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

GET

/v1/posts/{id}/status

Get a post (scheduled or created), including per-platform publish results

Key permission: posts

source is "scheduled" or "immediate" (POST /v1/posts/create). status is pending → publishing → published | failed, with "processing" while a platform finishes processing media. publish_results holds each platform's status (success | failed | processing), url when known, and error.

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/posts/d4e5f6a7-b8c9-4d0e-a1f2-3b4c5d6e7f8a/status" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

DELETE

/v1/posts/scheduled/{id}

Cancel a scheduled post

Key permission: posts

curl -X DELETE "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/posts/scheduled/d4e5f6a7-b8c9-4d0e-a1f2-3b4c5d6e7f8a" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

Projects

Separate sets of accounts under one account — one per client, or per project in an app you build on this API. A project holds posting accounts (X, TikTok, Instagram, Threads, Facebook Pages, LinkedIn), messaging channels (WhatsApp Business, Messenger, Instagram DMs, Telegram, Slack, Discord), a Google account (Gmail, Calendar, Drive) and Meta ad accounts (meta_ads, run with the facebook_ads tool and project_id on POST /v1/tools/facebook_ads). Posts and messages that name a project (project_id, or an X-Project-Id header on /v1/posts and /v1/platforms) use only that project's accounts and never fall back to yours. Your end users connect accounts in the hosted Connect portal (POST /v1/projects/{id}/connect-session) without a MultiverseAI login; inbound messages are pushed to your webhook (see Webhooks). {id} is our project id or your own external_id. /v1/profiles is an alias.

Connect portal popup: open the session url in a popup (e.g. window.open(url, "mv-connect", "width=520,height=720")). The portal lists the allowed platforms with Connect / Disconnect; provider sign-ins run in a child window and come back to the portal.
The popup posts events to your page: window.opener.postMessage({ source: 'multiverseai-connect', type: 'ready' | 'connected' | 'disconnected' | 'closed', platform, project_id, account }, return_origin). Check event.origin is https://multiverseai.com and event.data.source === 'multiverseai-connect'. return_origin is the session's origin, or return_url's origin; with neither, nothing is posted.
When the user clicks Done, a session with return_url sends the popup there with ?mv_status=done&mv_project_id=…; otherwise the popup posts closed and closes.
Account changes also arrive on your webhook as account.connected / account.disconnected, so you don't depend on the popup staying open.

POST

/v1/projects

Create a project

Key permission: posts

Request body

Field	Type	Required	Description
`name`	`string`	Yes	Display name
`external_id`	`string`	No	Your own id for it (unique per account). Creating again with the same external_id returns the existing project (200).
`metadata`	`object`	No	Anything you want stored with it

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/projects" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "name": "Acme Co",
  "external_id": "acme-client-42"
}'

GET

/v1/projects

List projects

Key permission: posts

Query parameters

Field	Type	Required	Description
`external_id`	`string`	No	Only the project with this external_id

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/projects" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

GET

/v1/projects/{id}

Get a project with its posting platforms, messaging/workspace channels, ads connections and connected accounts

Key permission: posts

ads lists the meta_ads connection(s): [{ name: "meta_ads", connected, account, accounts: [{ …account, status: "connected" | "expired", ad_accounts, pages }] }]. meta_ads is not a posting platform, so it is not under platforms or GET /v1/platforms.

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/projects/acme-client-42" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

PATCH

/v1/projects/{id}

Rename, or change external_id / metadata

Key permission: posts

curl -X PATCH "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/projects/acme-client-42" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "name": "Acme Corporation"
}'

DELETE

/v1/projects/{id}

Delete a project and disconnect its accounts

Key permission: posts

Its pending scheduled posts fail instead of posting as your account-level connections.

curl -X DELETE "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/projects/acme-client-42" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

POST

/v1/projects/{id}/connect-session

Create a hosted Connect portal session (popup)

Key permission: posts

The url carries a signed token scoped to this one project and these platforms; it expires at expires_at. Anyone holding it can connect or disconnect those platforms on this project until then — create one per user visit.
platforms[].available is false when a platform can't be connected yet (unavailable_code NOT_CONFIGURED, or PENDING_META_APPROVAL for WhatsApp Embedded Signup); the portal shows it greyed out.
Provider tokens never reach the portal page or your app.

Request body

Field	Type	Required	Description
`platforms`	`string[]`	No	Platforms to offer, in any order (default all): whatsapp, messenger, instagram_dm, telegram, slack, discord, google, facebook, instagram, threads, twitter (or x), linkedin, tiktok, youtube
`return_url`	`string`	No	https URL (http on localhost) the Done button sends the popup to, with ?mv_status=done&mv_project_id=…
`origin`	`string`	No	Origin the popup posts its events to (e.g. https://app.example.com). Default: return_url's origin
`theme`	`string`	No	light \| dark \| auto (default)
`partner_name`	`string`	No	Your product name, shown in the portal (60 characters max)
`ttl_seconds`	`number`	No	60–7200 (default 1800)

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/projects/acme-client-42/connect-session" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "platforms": [
    "whatsapp",
    "messenger",
    "instagram_dm",
    "telegram",
    "google"
  ],
  "origin": "https://app.example.com",
  "partner_name": "Acme",
  "theme": "auto"
}'

POST

/v1/projects/{id}/connect

Get a connect link for one platform (or connect a Telegram bot)

Key permission: posts

Send the account owner to url; they don't need a MultiverseAI login. The link is good for 10 minutes — make one per attempt. For several platforms at once, prefer the Connect portal (connect-session).
They come back to return_url with mv_status=connected|error, mv_platform, mv_project_id, mv_account, mv_account_id and mv_reason (and connected=<platform> on success).
Connecting again adds another account; Facebook / Messenger add every Page the person grants, instagram_dm every Instagram account linked to those Pages, WhatsApp every number of the shared WhatsApp Business Accounts.
WhatsApp uses Meta Embedded Signup. Until Meta approves it the call answers 503 PENDING_META_APPROVAL (or the person returns with mv_reason=whatsapp_pending_approval).
A Telegram bot has one webhook: connecting it to a project takes it over from anywhere it was paired before.
meta_ads connects the person's Meta ad accounts (Facebook Login for Business: ads_management, ads_read, business_management) for the facebook_ads tool; it is offered here only, not in the Connect portal. It keeps their long-lived user token (about 60 days; reconnect when the account shows status expired). On success the return URL also gets mv_ad_accounts=act_1,act_2…; mv_reason=ads_permission_declined when they declined ads access. Reconnecting the same Meta user updates the same account.

Request body

Field	Type	Required	Description
`platform`	`string`	Yes	twitter \| tiktok \| instagram \| threads \| facebook \| linkedin \| youtube \| whatsapp \| messenger \| instagram_dm \| slack \| discord \| google \| telegram \| meta_ads
`return_url`	`string`	No	Required except for telegram: https URL (http allowed on localhost) the person is sent back to
`bot_token`	`string`	No	telegram only: the bot token from @BotFather. The bot is connected immediately and its webhook pointed at MultiverseAI.
`google_services`	`string[]`	No	google only: gmail, calendar, drive (default all)

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/projects/acme-client-42/connect" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "platform": "instagram",
  "return_url": "https://app.example.com/social/connected"
}'

POST

/v1/projects/{id}/messages

Send a message through one of the project's channels

Key permission: posts

Uses only the project's own account for the platform — never yours.
Errors: 409 NOT_CONNECTED, 422 WINDOW_EXPIRED (Meta 24-hour window), 403 FORBIDDEN (Discord channel outside the project's server), 422 NOT_IN_CHANNEL (invite the Slack app first), 429 RATE_LIMITED, 422 PROVIDER_REJECTED.
Discord: messages go out as the MultiverseAI bot, only in the server connected to the project. Inbound Discord messages are not delivered (the bot has no gateway connection).

Request body

Field	Type	Required	Description
`platform`	`string`	Yes	whatsapp \| messenger \| instagram_dm (or instagram) \| telegram \| slack \| discord
`to`	`string`	Yes	WhatsApp phone number, Messenger PSID, Instagram IGSID, Telegram chat id, Slack channel id or Discord channel id — the chat_id of an inbound message.received event
`text`	`string`	No	Up to 4096 characters. Required unless media or template is sent.
`media`	`object`	No	{ kind: "image" \| "video" \| "audio" \| "document", url: "https://…", name? }
`template`	`object`	No	WhatsApp only: { name, language? (default en_US), components? } — needed to message someone outside the 24-hour window
`thread_ts`	`string`	No	Slack: reply in this thread
`account`	`string`	No	When the project has several accounts on the platform: our account id, the platform's id, or the username

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/projects/acme-client-42/messages" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "platform": "whatsapp",
  "to": "+15550100",
  "text": "Thanks — your order ships today."
}'

GET

/v1/projects/{id}/gmail/messages

List emails in the project's Gmail

Key permission: posts

Query parameters

Field	Type	Required	Description
`q`	`string`	No	Gmail search, e.g. is:unread from:alice@example.com
`limit`	`number`	No	1–50 (default 20)
`page_token`	`string`	No	next_page_token from the previous page
`account`	`string`	No	Which Google account when the project has several

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/projects/acme-client-42/gmail/messages?q=is:unread&limit=10" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

GET

/v1/projects/{id}/gmail/messages/{message_id}

Read one email (headers, text and HTML body)

Key permission: posts

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/projects/acme-client-42/gmail/messages/18c2a1b3c4d5e6f7" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

POST

/v1/projects/{id}/gmail/messages

Send an email from the project's Gmail

Key permission: posts

Request body

Field	Type	Required	Description
`to`	`string \| string[]`	Yes	Recipient address(es)
`subject`	`string`	Yes	Subject line
`text`	`string`	No	Plain-text body (text or html is required)
`html`	`string`	No	HTML body (with text, both are sent as alternatives)
`cc`	`string \| string[]`	No
`bcc`	`string \| string[]`	No
`thread_id`	`string`	No	Reply inside this Gmail thread
`in_reply_to`	`string`	No	Message-ID header of the email being answered

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/projects/acme-client-42/gmail/messages" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "to": "alice@example.com",
  "subject": "Your quote",
  "text": "Hi Alice, attached is…"
}'

GET

/v1/projects/{id}/calendar/events

List upcoming events in the project's Google Calendar

Key permission: posts

Query parameters

Field	Type	Required	Description
`time_min`	`string`	No	ISO 8601 (default now)
`time_max`	`string`	No	ISO 8601
`q`	`string`	No	Free-text search
`limit`	`number`	No	1–250 (default 25)
`calendar_id`	`string`	No	Default primary
`page_token`	`string`	No

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/projects/acme-client-42/calendar/events?time_max=2026-10-19T00:00:00Z" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

POST

/v1/projects/{id}/calendar/events

Create an event in the project's Google Calendar

Key permission: posts

Request body

Field	Type	Required	Description
`summary`	`string`	Yes	Title
`start`	`string`	Yes	RFC 3339 date-time, or YYYY-MM-DD for an all-day event
`end`	`string`	Yes	RFC 3339 date-time, or YYYY-MM-DD (exclusive) for all-day
`time_zone`	`string`	No	IANA zone for start/end, e.g. America/Los_Angeles
`description`	`string`	No
`location`	`string`	No
`attendees`	`string[]`	No	Email addresses
`google_meet`	`boolean`	No	Add a Google Meet link
`send_updates`	`string`	No	all \| externalOnly \| none (default none: attendees are not emailed)
`calendar_id`	`string`	No	Default primary

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/projects/acme-client-42/calendar/events" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "summary": "Kickoff",
  "start": "2026-10-14T17:00:00Z",
  "end": "2026-10-14T17:30:00Z",
  "attendees": [
    "alice@example.com"
  ],
  "google_meet": true
}'

DELETE

/v1/projects/{id}/calendar/events/{event_id}

Delete an event

Key permission: posts

Query parameters

Field	Type	Required	Description
`calendar_id`	`string`	No	Default primary

curl -X DELETE "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/projects/acme-client-42/calendar/events/abc123def456" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

GET

/v1/projects/{id}/drive/files

List or search files in the project's Google Drive

Key permission: posts

Read-only (drive.readonly scope). Newest first; trashed files are excluded.

Query parameters

Field	Type	Required	Description
`q`	`string`	No	Full-text search, or Drive query syntax (e.g. mimeType = 'application/pdf')
`limit`	`number`	No	1–100 (default 25)
`page_token`	`string`	No

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/projects/acme-client-42/drive/files?q=invoice" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

GET

/v1/projects/{id}/drive/files/{file_id}

Metadata of one Drive file (name, type, size, link)

Key permission: posts

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/projects/acme-client-42/drive/files/1AbCdEf…" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

GET

/v1/projects/{id}/accounts

List a project's connected accounts (no tokens)

Key permission: posts

meta_ads accounts also list ad_accounts (id, account_id, name, currency, timezone_name, account_status, pixels when readable) and pages (id, name, instagram_business_account), as saved at connect time.

Query parameters

Field	Type	Required	Description
`platform`	`string`	No	Only this platform

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/projects/acme-client-42/accounts?platform=meta_ads" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

GET

/v1/projects/{id}/insights

Live analytics for a project's connected accounts

Key permission: posts

One entry per connected posting account (instagram, facebook, threads, tiktok, twitter, linkedin; messaging channels, Google and meta_ads are not included), read live with that account's own connection. Unknown values are null, never 0. A platform that fails or lacks a permission says why in error; the other entries are unaffected and the request still returns 200.
daily has one row per day in the range. posts are the account's posts published in the range, newest first, up to 50. audience is null where the platform or connection gives none.
Instagram: followers, following, posts_count, post likes/comments always; daily reach and new_followers, post impressions (views), reach, shares, saves and audience need instagram_business_manage_insights. Facebook Pages: daily impressions, reach, engagements, profile_views, followers (as Meta still reports them) and post impressions/reach/reactions/comments/shares. Threads: daily impressions (views), followers, post views/likes/replies/reposts+quotes, audience (100+ followers). TikTok: profile and per-video views, likes, comments, shares; followers only with user.info.stats. X: profile counts and own tweets' public_metrics (tweets need an X API tier that allows reading them). LinkedIn: account only (no analytics scope).

Query parameters

Field	Type	Required	Description
`platform`	`string`	No	instagram \| facebook \| threads \| tiktok \| twitter \| linkedin (default: every connected account)
`since`	`string`	No	YYYY-MM-DD, inclusive (default: 29 days before until)
`until`	`string`	No	YYYY-MM-DD, inclusive (default: today, UTC). At most 90 days in all.
`refresh`	`boolean`	No	true skips the 10-minute cacheDefault: `false`

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/projects/acme-client-42/insights?platform=instagram&since=2026-10-01&until=2026-10-07" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

DELETE

/v1/projects/{id}/accounts/{account_id}

Disconnect one account

Key permission: posts

Telegram bots get their webhook removed, Slack and Google tokens are revoked; your webhook receives account.disconnected.

curl -X DELETE "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/projects/acme-client-42/accounts/9b2f…" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

Messaging (connected apps)

Read which WhatsApp number, Facebook Page and Instagram account a user connected, and reply from them. Partner apps such as Kanvo get a key limited to these endpoints when the user approves them at multiverseai.com/authorize; new messages are pushed to the app, signed, rather than polled.

GET

/v1/messaging/accounts

The connected messaging accounts

id is what Meta webhooks are keyed by: the phone_number_id, Page id or Instagram account id.

Query parameters

Field	Type	Required	Description
`project_id`	`string`	No	A project's channels instead (project id or external_id). Not available to connected-app keys.

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/messaging/accounts" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

POST

/v1/messaging/send

Send a message from a connected account

With media and text, WhatsApp sends the text as the caption (audio: as a second message); Messenger and Instagram send the file, then the text.
Errors: 409 NOT_CONNECTED, 422 WINDOW_EXPIRED (the 24-hour reply window has closed), 422 PROVIDER_REJECTED.

Request body

Field	Type	Required	Description
`platform`	`string`	Yes	"whatsapp", "messenger" or "instagram"
`to`	`string`	Yes	Phone number (WhatsApp) or the page-scoped id of the person (Messenger / Instagram)
`text`	`string`	No	Up to 4096 characters. Required unless media is sent.
`media`	`object`	No	{ kind: "image" \| "video" \| "audio" \| "document", url: "https://…", name? }
`project_id`	`string`	No	Send through a project's own channel instead (platform may then also be instagram_dm, telegram, slack or discord; template, thread_ts and account are accepted). Same as POST /v1/projects/{id}/messages. Not available to connected-app keys.

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/messaging/send" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "platform": "whatsapp",
  "to": "+15550100",
  "text": "Thanks — on its way!"
}'

POST

/v1/messaging/disconnect

Revoke the calling connected-app key

Stops event delivery to the app as well. The user can also disconnect from Settings → Developer.

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/messaging/disconnect" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{}'

Webhooks

Where MultiverseAI pushes events about your projects: messages people send to a project's WhatsApp number, Facebook Page, Instagram account, Telegram bot or Slack workspace, delivery statuses, accounts connected or disconnected (for example in the Connect portal), and how each post went on each platform. One webhook per API key. Every delivery is signed with the webhook's secret.

Delivery: POST <url> with Content-Type: application/json and headers X-Multiverse-Event (the event type), X-Multiverse-Delivery (the event id, the same on every retry — dedupe on it) and X-Multiverse-Signature: t=<unix seconds>,v1=<hex HMAC-SHA256 of "<t>.<raw body>" with your secret>. Reject a signature whose t is more than 5 minutes off.
Body: { id: "evt_…", object: "event", type, created_at, project_id, project: { id, external_id, name }, data }.
message.received data: { platform: whatsapp | messenger | instagram_dm | telegram | slack, account: { id, external_account_id, username }, message: { id, chat_id, from: { id, name, username? }, type, text, timestamp, attachments?, thread_ts? } }. Reply with POST /v1/projects/{id}/messages { platform, to: message.chat_id }.
message.status data: the same shape for WhatsApp sent / delivered / read / failed statuses (message.status, message.recipientId). account.connected / account.disconnected data: { platform, account: { id, external_account_id, username, display_name? } }.
post.published data: { post_id, platform, source: "immediate" | "scheduled", url, platform_post_id } — one per platform, when it is live (url is null when the platform gives no public link, e.g. a private TikTok). post.failed data: { post_id, platform, source, error }. They fire for created and scheduled posts, including Instagram/TikTok media that finished processing later; for a post without a project, project_id and project are null.
Answer with a 2xx within 10 seconds. Timeouts, network errors, 408, 429 and 5xx are retried twice (after about 2 s and 10 s); other 4xx are not retried. After 50 failed deliveries in a row the webhook is switched off — PUT it again to resume. Order is not guaranteed.
Discord messages and Gmail are not pushed: send to Discord with /v1/projects/{id}/messages and poll Gmail with /v1/projects/{id}/gmail/messages.
Provider tokens never appear in an event.

PUT

/v1/webhooks

Create or replace this key's webhook

Key permission: posts

The response includes secret (whsec_…) only when the webhook is created. Store it; replacing the URL keeps the same secret. POST is accepted as well as PUT.

Request body

Field	Type	Required	Description
`url`	`string`	Yes	Public https URL (no localhost or private addresses)
`events`	`string[]`	No	message.received, message.status, account.connected, account.disconnected, post.published, post.failed or "" (default [""])

curl -X PUT "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/webhooks" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "url": "https://app.example.com/multiverse/webhook",
  "events": [
    "message.received",
    "account.connected",
    "account.disconnected"
  ]
}'

GET

/v1/webhooks

This key's webhook and its last delivery status

Key permission: posts

A user session (not a key) sees every webhook on the account. The secret is never returned here.

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/webhooks" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

DELETE

/v1/webhooks

Remove this key's webhook

Key permission: posts

Deleting or revoking the API key stops delivery too.

curl -X DELETE "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/webhooks" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

POST

/v1/webhooks/rotate-secret

Issue a new signing secret

Key permission: posts

Returned once. The old secret stops working immediately, so deploy the new one right away.

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/webhooks/rotate-secret" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{}'

POST

/v1/webhooks/test

Send a ping event now

Key permission: posts

No retries. status is your endpoint's HTTP status (0 = timeout or network error).

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/webhooks/test" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{}'

Account & Usage

Who the caller is, your developer keys and their request log. Keys are read-only here: create and revoke them in Settings → Developer, so a leaked key can't mint new ones.

GET

/v1/me

Who you are, and which key you are calling with

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/me" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

GET

/v1/keys

List your developer keys (no secrets)

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/keys" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

GET

/v1/keys/{id}

Get one developer key

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/keys/2b3c4d5e-6f7a-4b8c-9d0e-1f2a3b4c5d6e" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

GET

/v1/usage

Request log for your keys, newest first

Query parameters

Field	Type	Required	Description
`key_id`	`uuid`	No	Only this key
`since`	`string`	No	ISO 8601; only requests at or after this time
`status`	`integer`	No	Only this HTTP status, e.g. 403
`limit`	`integer`	No	Page size, up to 500Default: `100`
`offset`	`integer`	No	Items to skipDefault: `0`

curl "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/public-api/v1/usage?key_id=2b3c4d5e-6f7a-4b8c-9d0e-1f2a3b4c5d6e&limit=2" \
  -H "Authorization: Bearer $MULTIVERSE_API_KEY"

Integration functions

The sections below document individual MultiverseAI functions that are called directly at https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/<function-name>, outside the versioned /v1 API. Their request and response formats are function-specific and do not follow the /v1 conventions above.

Web Scraping

Extract content from websites using Firecrawl, Browserless, or ScrapingBee. Supports user-provided API keys.

POST

/web-scraping-api

Unified web scraping API - intelligently routes to Firecrawl, Browserless, or ScrapingBee based on operation and configuration. Supports user-provided API keys.

Request body

Field	Type	Required	Description
`url`	`string`	Yes	The URL to scrape
`operation`	`string`	No	Operation: scrape, screenshot, pdf, search, map, crawlDefault: `scrape`
`provider`	`string`	No	Provider: auto, firecrawl, browserless, scrapingbeeDefault: `auto`
`apiKeys`	`object`	No	User API keys: { firecrawl?, browserless?, scrapingbee? }
`formats`	`array`	No	Output formats (Firecrawl): markdown, html, links, screenshot
`selector`	`string`	No	CSS selector (Browserless/ScrapingBee)
`executeJs`	`boolean`	No	Execute JavaScript on page
`jsScript`	`string`	No	Custom JavaScript to execute
`useProxy`	`boolean`	No	Enable proxy rotation (ScrapingBee)
`renderJs`	`boolean`	No	Render JavaScript (ScrapingBee)

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/web-scraping-api" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://example.com",
    "operation": "scrape",
    "provider": "auto",
    "apiKeys": {
      "firecrawl": "fc-your-key",
      "browserless": "your-browserless-key"
    }
  }'

POST

/browserless-scrape

Browserless.io scraping - content extraction, screenshots, PDFs, and custom JavaScript execution

Request body

Field	Type	Required	Description
`url`	`string`	Yes	The URL to scrape
`apiKey`	`string`	No	Your Browserless API key (optional, uses system key if not provided)
`operation`	`string`	No	Operation: content, screenshot, pdf, functionDefault: `content`
`selector`	`string`	No	CSS selector to target specific elements
`waitForSelector`	`string`	No	Wait for element to appear
`timeout`	`integer`	No	Timeout in millisecondsDefault: `30000`
`fullPage`	`boolean`	No	Capture full page screenshotDefault: `false`
`executeJs`	`boolean`	No	Execute custom JavaScriptDefault: `false`
`jsScript`	`string`	No	JavaScript code to execute on page

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/browserless-scrape" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://example.com",
    "operation": "screenshot",
    "fullPage": true,
    "apiKey": "your-browserless-api-key"
  }'

POST

/scrapingbee-scrape

ScrapingBee scraping - proxy rotation, JavaScript rendering, and CSS extraction

Request body

Field	Type	Required	Description
`url`	`string`	Yes	The URL to scrape
`apiKey`	`string`	No	Your ScrapingBee API key (optional, uses system key if not provided)
`renderJs`	`boolean`	No	Render JavaScript (uses more credits)Default: `false`
`premiumProxy`	`boolean`	No	Use premium proxy poolDefault: `false`
`countryCode`	`string`	No	Proxy country code (e.g., "us", "uk")
`extractRules`	`object`	No	CSS extraction rules: { title: "h1", links: { selector: "a", type: "list" } }
`waitFor`	`string`	No	CSS selector to wait for
`timeout`	`integer`	No	Timeout in millisecondsDefault: `30000`
`returnScreenshot`	`boolean`	No	Return page screenshotDefault: `false`

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/scrapingbee-scrape" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://example.com",
    "renderJs": true,
    "extractRules": {
      "title": "h1",
      "paragraphs": { "selector": "p", "type": "list" }
    },
    "apiKey": "your-scrapingbee-api-key"
  }'

POST

/firecrawl-scrape

Scrape a single URL and extract content in various formats (markdown, HTML, screenshot, etc.)

Request body

Field	Type	Required	Description
`url`	`string`	Yes	The URL to scrape
`formats`	`array`	No	Output formats: markdown, html, rawHtml, links, screenshot, branding, summaryDefault: `["markdown"]`
`onlyMainContent`	`boolean`	No	Exclude headers, navs, footersDefault: `true`
`waitFor`	`integer`	No	Milliseconds to wait for dynamic content
`location`	`object`	No	Geo-targeting: { country: "US", languages: ["en"] }

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/firecrawl-scrape" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://example.com",
    "formats": ["markdown", "links"],
    "onlyMainContent": true
  }'

POST

/firecrawl-search

Search the web and optionally scrape content from results

Request body

Field	Type	Required	Description
`query`	`string`	Yes	Search query string
`limit`	`integer`	No	Number of results (1-100)Default: `10`
`lang`	`string`	No	Language code (e.g., "en")
`country`	`string`	No	Country code (e.g., "US")
`tbs`	`string`	No	Time filter: qdr:h (hour), qdr:d (day), qdr:w (week), qdr:m (month), qdr:y (year)
`scrapeOptions`	`object`	No	If provided, returns full content from each result

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/firecrawl-search" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "latest AI news",
    "limit": 5,
    "tbs": "qdr:w"
  }'

POST

/firecrawl-map

Quickly discover all URLs on a website (fast sitemap generation)

Request body

Field	Type	Required	Description
`url`	`string`	Yes	The base URL to map
`search`	`string`	No	Filter URLs by keyword (ordered by relevance)
`limit`	`integer`	No	Maximum URLs to returnDefault: `5000`
`includeSubdomains`	`boolean`	No	Include subdomain URLsDefault: `false`

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/firecrawl-map" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://example.com",
    "limit": 100
  }'

POST

/firecrawl-crawl

Recursively crawl and scrape all pages on a website

Request body

Field	Type	Required	Description
`url`	`string`	Yes	The starting URL to crawl
`limit`	`integer`	No	Maximum pages to crawlDefault: `100`
`maxDepth`	`integer`	No	Maximum link depth to follow
`includePaths`	`array`	No	URL patterns to include
`excludePaths`	`array`	No	URL patterns to exclude

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/firecrawl-crawl" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://example.com",
    "limit": 50,
    "maxDepth": 2
  }'

POST

/webscraper

Puppeteer-based web scraping with JavaScript execution support

Request body

Field	Type	Required	Description
`url`	`string`	Yes	The URL to scrape
`selector`	`string`	No	CSS selector to target specific elements
`scrapeType`	`string`	No	Output type: text, html, or screenshotDefault: `text`
`waitForSelector`	`string`	No	Wait for specific element to appear
`timeout`	`integer`	No	Timeout in millisecondsDefault: `30000`
`executeJavaScript`	`boolean`	No	Execute custom JavaScriptDefault: `false`
`jsScript`	`string`	No	JavaScript code to execute on page

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/webscraper" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://example.com",
    "scrapeType": "text",
    "selector": "article"
  }'

Search

AI-powered web search with citations using Perplexity

POST

/perplexity-search

Perform AI-powered web search with grounded responses and citations

Request body

Field	Type	Required	Description
`query`	`string`	Yes	The search query
`model`	`string`	No	Model: sonar, sonar-pro, sonar-reasoning, sonar-reasoning-pro, sonar-deep-researchDefault: `sonar`
`systemPrompt`	`string`	No	System prompt for the AIDefault: `Be precise and concise.`
`maxTokens`	`integer`	No	Maximum response tokens
`temperature`	`number`	No	Sampling temperature (0-2)Default: `0.2`
`searchDomainFilter`	`array`	No	Domains to include/exclude (prefix with - to exclude)
`searchRecencyFilter`	`string`	No	Time filter: day, week, month, year
`searchMode`	`string`	No	Search mode: web (default), academic, sec

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/perplexity-search" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "What are the latest developments in quantum computing?",
    "model": "sonar-pro",
    "searchRecencyFilter": "week"
  }'

Semantic Search

AI-powered semantic search with multiple search modes and vector embeddings for content retrieval

POST

/semantic-search

Perform semantic similarity, hybrid, fuzzy, keyword, or multimodal search over indexed content using vector embeddings

Key permission: read

Request body

Field	Type	Required	Description
`query`	`string`	Yes	Natural language search query
`searchType`	`string`	No	Search mode: similarity, hybrid, fuzzy, keyword, multimodalDefault: `similarity`
`vectorProvider`	`string`	No	Vector DB providerDefault: `pinecone`
`embeddingProvider`	`string`	No	Embedding providerDefault: `openai`
`embeddingModel`	`string`	No	Embedding model to useDefault: `text-embedding-3-small`
`topK`	`integer`	No	Maximum results to returnDefault: `10`
`similarityThreshold`	`number`	No	Minimum similarity score (0-1)Default: `0.7`
`dimensions`	`integer`	No	Embedding dimensionsDefault: `1536`
`rerankProvider`	`string`	No	Optional reranking providerDefault: `none`
`hybridWeights`	`object`	No	Weights for hybrid search: { vector: 0.7, keyword: 0.3 }
`filters`	`array`	No	Array of filter objects: [{ field, operator, value }]
`boosts`	`array`	No	Array of boost objects: [{ field, value, weight }]
`includeMetadata`	`boolean`	No	Include metadata in resultsDefault: `true`
`includeScore`	`boolean`	No	Include similarity scoresDefault: `true`
`advancedOptions`	`object`	No	{ fuzzyThreshold, expandQuery, spellCheck, diversity }

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/semantic-search" \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "How to set up OAuth integration",
    "searchType": "hybrid",
    "topK": 5,
    "similarityThreshold": 0.75,
    "hybridWeights": { "vector": 0.7, "keyword": 0.3 }
  }'

POST

/ai-embeddings

Generate vector embeddings for text content. Supports OpenAI, Azure, Cohere, HuggingFace, Google, and custom providers.

Key permission: write

Request body

Field	Type	Required	Description
`texts`	`array`	Yes	Array of text strings to generate embeddings for
`provider`	`string`	No	Embedding provider: openai, azure, cohere, huggingface, google, custom
`model`	`string`	No	Model name (e.g., text-embedding-3-small)
`dimensions`	`integer`	No	Output embedding dimensionsDefault: `1536`
`batchSize`	`integer`	No	Batch size for processing multiple textsDefault: `100`
`maxLength`	`integer`	No	Max text length before truncationDefault: `512`
`normalization`	`boolean`	No	Normalize vectors to unit lengthDefault: `true`
`truncation`	`boolean`	No	Truncate texts exceeding maxLengthDefault: `true`
`customEndpoint`	`string`	No	Custom API endpoint URL (for azure or custom providers)

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/ai-embeddings" \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "texts": ["Quantum computing uses quantum bits", "Machine learning models learn from data"],
    "provider": "openai",
    "model": "text-embedding-3-small",
    "dimensions": 1536
  }'

YouTube

Access YouTube Data API for videos, channels, and playlists

POST

/youtube-api

Proxy endpoint for YouTube Data API v3. Requires API key authentication.

Key permission: read

Request body

Field	Type	Required	Description
`endpoint`	`string`	Yes	YouTube API endpoint path (e.g., /search, /videos, /channels)
`params`	`object`	No	Query parameters for the YouTube API

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/youtube-api" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "endpoint": "/search",
    "params": {
      "part": "snippet",
      "q": "programming tutorials",
      "type": "video",
      "maxResults": 10
    }
  }'

QuickTube

YouTube video analysis powered by QuickTube AI - transcripts, summaries, best sections, and more

POST

/quicktube-api

QuickTube AI API - transcribe, summarize, analyze, and extract insights from YouTube videos

Request body

Field	Type	Required	Description
`action`	`string`	Yes	Action: transcript, summarize, best-sections, characters, chat, quiz, speech, download-video, video-metadata, video-search, video-comments, batch-process
`url`	`string`	No	YouTube video URL (for transcript, summarize, best-sections, characters, chat, quiz)
`video_id`	`string`	No	YouTube video ID (for metadata, comments, related)
`query`	`string`	No	Search query (for video-search) or chat question (for chat)
`language`	`string`	No	Language code for transcriptsDefault: `en`

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/quicktube-api" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "action": "summarize",
    "url": "https://www.youtube.com/watch?v=dQw4w9WgXcQ"
  }'

POST

/quicktube-api (transcript)

Get a full transcript of a YouTube video with timestamps

Request body

Field	Type	Required	Description
`action`	`string`	Yes	Set to "transcript"
`url`	`string`	Yes	YouTube video URL
`language`	`string`	No	Language codeDefault: `en`

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/quicktube-api" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "action": "transcript",
    "url": "https://www.youtube.com/watch?v=VIDEO_ID"
  }'

POST

/quicktube-api (best-sections)

AI-powered extraction of the most engaging video segments

Request body

Field	Type	Required	Description
`action`	`string`	Yes	Set to "best-sections"
`url`	`string`	Yes	YouTube video URL

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/quicktube-api" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "action": "best-sections",
    "url": "https://www.youtube.com/watch?v=VIDEO_ID"
  }'

Integrations

Send emails and Slack messages through our integration endpoints

POST

/send-email

Send emails via Resend API

Request body

Field	Type	Required	Description
`to`	`string \| array`	Yes	Recipient email(s)
`subject`	`string`	Yes	Email subject line
`body`	`string`	Yes	Email content (text or HTML)
`html`	`boolean`	No	Set to true if body contains HTMLDefault: `false`
`from_name`	`string`	No	Sender display nameDefault: `MultiverseAI`
`reply_to`	`string`	No	Reply-to email address

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/send-email" \
  -H "Content-Type: application/json" \
  -d '{
    "to": "user@example.com",
    "subject": "Welcome to MultiverseAI",
    "body": "<h1>Welcome!</h1><p>Thanks for signing up.</p>",
    "html": true
  }'

POST

/slack-webhook

Send messages to Slack channels via webhook

Request body

Field	Type	Required	Description
`webhook_url`	`string`	Yes	Slack incoming webhook URL
`message`	`string`	No	Message text (required if no attachments)
`channel`	`string`	No	Override default channel
`username`	`string`	No	Override bot username
`icon_emoji`	`string`	No	Override bot icon (e.g., :robot_face:)
`attachments`	`array`	No	Rich message attachments with colors, fields, etc.

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/slack-webhook" \
  -H "Content-Type: application/json" \
  -d '{
    "webhook_url": "https://hooks.slack.com/services/T.../B.../xxx",
    "message": "New deployment completed! :rocket:",
    "username": "Deploy Bot",
    "icon_emoji": ":ship:",
    "attachments": [{
      "color": "#36a64f",
      "title": "Production Deploy",
      "fields": [
        { "title": "Version", "value": "v2.1.0", "short": true },
        { "title": "Status", "value": "Success", "short": true }
      ]
    }]
  }'

GitHub

GitHub App integration for repository access, issues, and pull requests

POST

/github-api

GitHub App API - list installations, repositories, and perform actions

Request body

Field	Type	Required	Description
`action`	`string`	Yes	Action: list-installations, list-repos, get-contents, get-file, get-issues, create-issue, create-branch, create-file, create-pr, add-comment, import-repo-context
`installationId`	`number`	No	GitHub App installation ID (required for most actions)
`owner`	`string`	No	Repository owner (required for repo-specific actions)
`repo`	`string`	No	Repository name (required for repo-specific actions)

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/github-api" \
  -H "Content-Type: application/json" \
  -d '{
    "action": "list-installations"
  }'

POST

/github-api (list-repos)

List repositories accessible by a GitHub App installation

Request body

Field	Type	Required	Description
`action`	`string`	Yes	Set to "list-repos"
`installationId`	`number`	Yes	GitHub App installation ID

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/github-api" \
  -H "Content-Type: application/json" \
  -d '{
    "action": "list-repos",
    "installationId": 12345678
  }'

POST

/github-api (create-issue)

Create a new issue in a GitHub repository

Request body

Field	Type	Required	Description
`action`	`string`	Yes	Set to "create-issue"
`installationId`	`number`	Yes	GitHub App installation ID
`owner`	`string`	Yes	Repository owner
`repo`	`string`	Yes	Repository name
`title`	`string`	Yes	Issue title
`body`	`string`	No	Issue body (markdown supported)
`labels`	`array`	No	Array of label names
`assignees`	`array`	No	Array of GitHub usernames to assign

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/github-api" \
  -H "Content-Type: application/json" \
  -d '{
    "action": "create-issue",
    "installationId": 12345678,
    "owner": "your-org",
    "repo": "my-project",
    "title": "Bug: Login not working",
    "body": "Users are unable to login...",
    "labels": ["bug", "high-priority"]
  }'

POST

/github-api (create-pr)

Create a pull request with code changes

Request body

Field	Type	Required	Description
`action`	`string`	Yes	Set to "create-pr"
`installationId`	`number`	Yes	GitHub App installation ID
`owner`	`string`	Yes	Repository owner
`repo`	`string`	Yes	Repository name
`title`	`string`	Yes	PR title
`body`	`string`	No	PR description
`head`	`string`	Yes	Branch with changes
`base`	`string`	No	Target branchDefault: `main`

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/github-api" \
  -H "Content-Type: application/json" \
  -d '{
    "action": "create-pr",
    "installationId": 12345678,
    "owner": "your-org",
    "repo": "my-project",
    "title": "Add new feature",
    "body": "This PR adds...",
    "head": "feature/new-feature",
    "base": "main"
  }'

POST

/github-api (import-repo-context)

Import repository code as AI context (useful for code analysis)

Request body

Field	Type	Required	Description
`action`	`string`	Yes	Set to "import-repo-context"
`installationId`	`number`	Yes	GitHub App installation ID
`owner`	`string`	Yes	Repository owner
`repo`	`string`	Yes	Repository name
`path`	`string`	No	Subdirectory to import (optional)

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/github-api" \
  -H "Content-Type: application/json" \
  -d '{
    "action": "import-repo-context",
    "installationId": 12345678,
    "owner": "your-org",
    "repo": "my-project",
    "path": "src"
  }'

Discord

Discord bot integration for sending messages, managing webhooks, and slash commands

POST

/discord-api

Discord API - send messages, manage webhooks, and interact with Discord servers

Request body

Field	Type	Required	Description
`action`	`string`	Yes	Action: get-config, list-guilds, get-guild, list-channels, send-message, send-embed, create-webhook, execute-webhook, register-commands, list-commands
`channelId`	`string`	No	Discord channel ID (for message/webhook actions)
`guildId`	`string`	No	Discord guild (server) ID
`content`	`string`	No	Message content

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/discord-api" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "action": "get-config"
  }'

POST

/discord-api (send-message)

Send a message to a Discord channel

Request body

Field	Type	Required	Description
`action`	`string`	Yes	Set to "send-message"
`channelId`	`string`	Yes	Discord channel ID
`content`	`string`	Yes	Message content

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/discord-api" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "action": "send-message",
    "channelId": "1234567890123456789",
    "content": "Hello from MultiverseAI!"
  }'

POST

/discord-api (send-embed)

Send a rich embed message to a Discord channel

Request body

Field	Type	Required	Description
`action`	`string`	Yes	Set to "send-embed"
`channelId`	`string`	Yes	Discord channel ID
`title`	`string`	Yes	Embed title
`description`	`string`	No	Embed description
`color`	`number`	No	Embed color (decimal, default: 5793266)
`fields`	`array`	No	Array of field objects
`footer`	`string`	No	Footer text

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/discord-api" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "action": "send-embed",
    "channelId": "1234567890123456789",
    "title": "Workflow Completed",
    "description": "Your AI workflow has finished",
    "color": 3066993,
    "footer": "MultiverseAI"
  }'

POST

/discord-api (execute-webhook)

Send a message via Discord webhook

Request body

Field	Type	Required	Description
`action`	`string`	Yes	Set to "execute-webhook"
`webhookId`	`string`	Yes	Discord webhook ID
`webhookToken`	`string`	Yes	Discord webhook token
`content`	`string`	No	Message content
`username`	`string`	No	Override webhook username

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/discord-api" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "action": "execute-webhook",
    "webhookId": "1234567890123456789",
    "webhookToken": "abc123...",
    "content": "New event triggered!",
    "username": "Event Bot"
  }'

POST

/discord-webhook

Webhook endpoint for Discord interactions (slash commands, buttons). Configure this URL in Discord Developer Portal.

Request body

Field	Type	Required	Description
`type`	`number`	No	Interaction type (1=PING, 2=APPLICATION_COMMAND)
`data`	`object`	No	Interaction data from Discord

# Discord sends interactions to this endpoint:
https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/discord-webhook

# Set this as your Interactions Endpoint URL in Discord Developer Portal

Gmail

Send and read emails via Gmail API with OAuth2 authentication

POST

/gmail-api

Gmail API - send emails, read inbox, manage labels. Requires Google OAuth connection.

Key permission: integrations

Request body

Field	Type	Required	Description
`operation`	`string`	Yes	Operation: send, list, get, search, labels, draft
`userId`	`string`	Yes	User ID (must have connected Google account)
`to`	`string`	No	Recipient email (for send operation)
`subject`	`string`	No	Email subject (for send operation)
`body`	`string`	No	Email body (for send operation)
`cc`	`string`	No	CC recipients (for send/draft)
`bcc`	`string`	No	BCC recipients (for send/draft)
`isHtml`	`boolean`	No	Set to true if body contains HTML (for send/draft)Default: `false`
`messageId`	`string`	No	Message ID (for get operation)
`query`	`string`	No	Gmail search query (for search operation)
`maxResults`	`integer`	No	Max messages to return (for list/search)Default: `20`
`labelIds`	`array`	No	Filter by label IDs, e.g. ["INBOX", "UNREAD"] (for list)
`includeBody`	`boolean`	No	Fetch full message bodies (for list/search)
`format`	`string`	No	Message format: full or minimal (for get)Default: `full`

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/gmail-api" \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "operation": "send",
    "userId": "user_abc123",
    "to": "recipient@example.com",
    "subject": "Hello from MultiverseAI",
    "body": "This email was sent via the Gmail integration."
  }'

POST

/gmail-api (list)

List inbox messages with optional label and search filtering

Key permission: integrations

Request body

Field	Type	Required	Description
`operation`	`string`	Yes	Set to "list"
`userId`	`string`	Yes	User ID
`maxResults`	`integer`	No	Maximum messages to returnDefault: `20`
`labelIds`	`array`	No	Filter by label IDs (e.g., ["INBOX", "UNREAD"])

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/gmail-api" \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "operation": "list",
    "userId": "user_abc123",
    "maxResults": 10
  }'

Google Drive

Upload, download, and manage files in Google Drive via OAuth2. Supports 12 operations including search, copy, move, and sharing.

POST

/google-drive-api

Google Drive API - list, search, get, create, update, delete, download, upload, copy, move, create_folder, and share files. Requires Google OAuth.

Key permission: integrations

Request body

Field	Type	Required	Description
`operation`	`string`	Yes	Operation: list, search, get, create, update, delete, download, upload, copy, move, create_folder, share
`userId`	`string`	Yes	User ID (must have connected Google account)
`fileId`	`string`	No	File ID (required for get, update, delete, download, copy, move, share)
`folderId`	`string`	No	Parent folder ID (for list, upload, create_folder)
`name`	`string`	No	File/folder name (for create, upload, copy, create_folder)
`content`	`string`	No	File content (for create, upload, update)
`mimeType`	`string`	No	MIME type (for create, upload)Default: `application/octet-stream`
`description`	`string`	No	File description (for create, update)
`query`	`string`	No	Search query (for list, search)
`pageSize`	`integer`	No	Max results per page (for list, search)Default: `10`
`pageToken`	`string`	No	Pagination token for next page
`shareEmail`	`string`	No	Email to share with (required for share)
`shareRole`	`string`	No	Share permission: reader, writer, commenterDefault: `reader`
`destinationFolderId`	`string`	No	Target folder ID (required for move)
`exportMimeType`	`string`	No	Export format for Google Docs files (for download)

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/google-drive-api" \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "operation": "list",
    "userId": "user_abc123",
    "folderId": "root",
    "pageSize": 20
  }'

POST

/google-drive-api (upload)

Upload a file to Google Drive with optional folder placement. Supports multipart uploads.

Key permission: integrations

Request body

Field	Type	Required	Description
`operation`	`string`	Yes	Set to "upload"
`userId`	`string`	Yes	User ID
`fileName`	`string`	Yes	File name
`content`	`string`	Yes	File content
`mimeType`	`string`	No	MIME type
`folderId`	`string`	No	Parent folder ID

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/google-drive-api" \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "operation": "upload",
    "userId": "user_abc123",
    "fileName": "report.txt",
    "content": "Report content here...",
    "mimeType": "text/plain",
    "folderId": "folder_123"
  }'

POST

/google-drive-api (share)

Share a file or folder with another user via email

Key permission: integrations

Request body

Field	Type	Required	Description
`operation`	`string`	Yes	Set to "share"
`userId`	`string`	Yes	User ID
`fileId`	`string`	Yes	File or folder ID to share
`shareEmail`	`string`	Yes	Recipient email address
`shareRole`	`string`	No	Permission level: reader, writer, commenterDefault: `reader`

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/google-drive-api" \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "operation": "share",
    "userId": "user_abc123",
    "fileId": "file_abc123",
    "shareEmail": "colleague@example.com",
    "shareRole": "writer"
  }'

Google Docs

Create, read, update, and export Google Docs via OAuth2

POST

/google-docs-api

Google Docs API - create, get, get_content, update, append, and export documents. Requires Google OAuth connection.

Key permission: integrations

Request body

Field	Type	Required	Description
`operation`	`string`	Yes	Operation: create, get, get_content, update, append, export
`userId`	`string`	Yes	User ID (must have connected Google account)
`documentId`	`string`	No	Document ID (required for get, get_content, update, append, export)
`title`	`string`	No	Document title (required for create)
`content`	`string`	No	Document content (required for create, update, append)
`folderId`	`string`	No	Google Drive folder ID to place document in (for create)
`insertLocation`	`string`	No	Where to insert content: start, end, or index (for update, default: end)
`exportFormat`	`string`	No	Export format: pdf, docx, txt, html, odt, rtf (for export, default: pdf)

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/google-docs-api" \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "operation": "create",
    "userId": "user_abc123",
    "title": "Meeting Notes",
    "content": "# Meeting Notes\n\n- Item 1\n- Item 2",
    "folderId": "folder_optional_123"
  }'

POST

/google-docs-api (get_content)

Extract plain text from a Google Doc, including content from paragraphs and tables

Key permission: integrations

Request body

Field	Type	Required	Description
`operation`	`string`	Yes	Set to "get_content"
`userId`	`string`	Yes	User ID
`documentId`	`string`	Yes	Google Docs document ID

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/google-docs-api" \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "operation": "get_content",
    "userId": "user_abc123",
    "documentId": "1BxiMVs0XRA5nFMdKvBdBZjgmUii7Op3Lr32456"
  }'

POST

/google-docs-api (append)

Append content to the end of a Google Doc with automatic newline separator

Key permission: integrations

Request body

Field	Type	Required	Description
`operation`	`string`	Yes	Set to "append"
`userId`	`string`	Yes	User ID
`documentId`	`string`	Yes	Google Docs document ID
`content`	`string`	Yes	Content to append

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/google-docs-api" \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "operation": "append",
    "userId": "user_abc123",
    "documentId": "1BxiMVs0XRA5nFMdKvBdBZjgmUii7Op3Lr32456",
    "content": "\nNew section added via API"
  }'

POST

/google-docs-api (export)

Export a Google Doc to PDF, DOCX, TXT, HTML, ODT, or RTF. Binary formats are returned as base64.

Key permission: integrations

Request body

Field	Type	Required	Description
`operation`	`string`	Yes	Set to "export"
`userId`	`string`	Yes	User ID
`documentId`	`string`	Yes	Google Docs document ID
`exportFormat`	`string`	No	Format: pdf, docx, txt, html, odt, rtfDefault: `pdf`

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/google-docs-api" \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "operation": "export",
    "userId": "user_abc123",
    "documentId": "1BxiMVs0XRA5nFMdKvBdBZjgmUii7Op3Lr32456",
    "exportFormat": "pdf"
  }'

POST

/google-docs-export

Create a formatted Google Doc with styled headers and sections. Applies bold formatting and font sizing automatically.

Key permission: integrations

Request body

Field	Type	Required	Description
`userId`	`string`	Yes	User ID
`title`	`string`	Yes	Document title
`content`	`string`	Yes	Document content (sections split by newlines, headers auto-detected)
`accessToken`	`string`	No	Optional Google access token (fetched from DB if not provided)

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/google-docs-export" \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "userId": "user_abc123",
    "title": "Quarterly Report",
    "content": "Revenue Summary\n---\nQ1 revenue grew 15% YoY..."
  }'

Google Workspace

Sheets, Slides, Calendar, Forms, Tasks, Contacts, Chat and Meet as the caller's connected Google account

POST

/google-workspace-api (sheets)

Google Sheets: read, append, write, findRows, clear, create, addSheet, get. Runs as the Google account connected in Settings → Integrations.

Key permission: integrations

Request body

Field	Type	Required	Description
`service`	`string`	Yes	"sheets"
`operation`	`string`	Yes	read (Values in a range, plus rows keyed by the header row); append (Adds rows after the last row of a table); write (Overwrites the cells of a range); findRows (Rows whose column (by header name) equals a value); clear (Clears values (keeps formatting)); create (A new spreadsheet, optionally with tabs and a header row); addSheet (Adds a sheet (tab) to a spreadsheet); get (Title, URL and tabs)
`params`	`object`	Yes	read: spreadsheetId, range, headerRow · append: spreadsheetId, range, values, valueInputOption · write: spreadsheetId, range, values, valueInputOption · findRows: spreadsheetId, range, column, value, maxResults · clear: spreadsheetId, range · create: title, sheetTitles, headers · addSheet: spreadsheetId, title · get: spreadsheetId*

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/google-workspace-api" \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"service": "sheets", "operation": "read", "params": {"spreadsheetId":"<spreadsheet>"}}'

POST

/google-workspace-api (slides)

Google Slides: create, addSlide, replaceText, get. Runs as the Google account connected in Settings → Integrations.

Key permission: integrations

Request body

Field	Type	Required	Description
`service`	`string`	Yes	"slides"
`operation`	`string`	Yes	create (A new, empty deck); addSlide (A slide with a title and body text); replaceText (Fills a template deck: every match in every slide); get (Title, URL and each slide's text)
`params`	`object`	Yes	create: title · addSlide: presentationId, title, body, layout, insertionIndex · replaceText: presentationId, replacements, find, replace, matchCase · get: presentationId*

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/google-workspace-api" \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"service": "slides", "operation": "create", "params": {}}'

POST

/google-workspace-api (calendar)

Google Calendar: createEvent, listEvents, getEvent, updateEvent, deleteEvent, quickAdd, freeBusy, listCalendars. Runs as the Google account connected in Settings → Integrations.

Key permission: integrations

Request body

Field	Type	Required	Description
`service`	`string`	Yes	"calendar"
`operation`	`string`	Yes	createEvent (Invites guests; can add a Meet link); listEvents (Upcoming events (from now unless set)); getEvent (One event by ID); updateEvent (Changes only the fields you fill in); deleteEvent (Cancels the event); quickAdd (An event from a sentence); freeBusy (Busy blocks per calendar); listCalendars (Calendars you can see, with their IDs)
`params`	`object`	Yes	createEvent: calendarId, summary, start, end, durationMinutes, timeZone, attendees, location, description, addMeet, sendUpdates · listEvents: calendarId, timeMin, timeMax, query, maxResults · getEvent: calendarId, eventId* · updateEvent: eventId, calendarId, summary, start, end, durationMinutes, timeZone, attendees, location, description, addMeet, sendUpdates · deleteEvent: calendarId, eventId, sendUpdates · quickAdd: calendarId, text* · freeBusy: calendars, timeMin, timeMax, timeZone · listCalendars: none

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/google-workspace-api" \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"service": "calendar", "operation": "createEvent", "params": {"summary":"<title>","start":"<start>"}}'

POST

/google-workspace-api (forms)

Google Forms: listResponses, getResponse, create, addQuestion, get. Runs as the Google account connected in Settings → Integrations.

Key permission: integrations

Request body

Field	Type	Required	Description
`service`	`string`	Yes	"forms"
`operation`	`string`	Yes	listResponses (Answers keyed by question title); getResponse (One response by ID); create (A new form; add questions with Add question); addQuestion (Appends a question); get (Title, links and questions)
`params`	`object`	Yes	listResponses: formId, since, maxResults · getResponse: formId, responseId* · create: title, description · addQuestion: formId, title, questionType, options, required · get: formId

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/google-workspace-api" \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"service": "forms", "operation": "listResponses", "params": {"formId":"<form>"}}'

POST

/google-workspace-api (tasks)

Google Tasks: createTask, listTasks, completeTask, updateTask, deleteTask, listTaskLists. Runs as the Google account connected in Settings → Integrations.

Key permission: integrations

Request body

Field	Type	Required	Description
`service`	`string`	Yes	"tasks"
`operation`	`string`	Yes	createTask (Adds a task to a list); listTasks (Open tasks (and completed ones if asked)); completeTask (Marks a task done); updateTask (Changes only the fields you fill in); deleteTask (Removes a task); listTaskLists (Your lists, with their IDs)
`params`	`object`	Yes	createTask: taskListId, title, notes, due · listTasks: taskListId, showCompleted, dueMax, maxResults · completeTask: taskListId, taskId · updateTask: taskListId, taskId, title, notes, due · deleteTask: taskListId, taskId · listTaskLists: none

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/google-workspace-api" \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"service": "tasks", "operation": "createTask", "params": {"title":"<title>"}}'

POST

/google-workspace-api (contacts)

Google Contacts: search, create, update, get, list, delete. Runs as the Google account connected in Settings → Integrations.

Key permission: integrations

Request body

Field	Type	Required	Description
`service`	`string`	Yes	"contacts"
`operation`	`string`	Yes	search (By name, email or phone); create (Adds a contact); update (Replaces the fields you fill in); get (One contact); list (Most recently changed first); delete (Removes a contact)
`params`	`object`	Yes	search: query, maxResults · create: givenName, familyName, email, phone, organization, title, notes · update: resourceName, givenName, familyName, email, phone, organization, title, notes · get: resourceName* · list: maxResults · delete: resourceName*

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/google-workspace-api" \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"service": "contacts", "operation": "search", "params": {"query":"<search>"}}'

POST

/google-workspace-api (chat)

Google Chat: sendMessage, listSpaces, listMessages, updateMessage, deleteMessage. Runs as the Google account connected in Settings → Integrations.

Key permission: integrations

Request body

Field	Type	Required	Description
`service`	`string`	Yes	"chat"
`operation`	`string`	Yes	sendMessage (Posts to a space as you); listSpaces (Spaces and DMs you're in, with their names); listMessages (Latest messages in a space); updateMessage (Replaces the text of your message); deleteMessage (Deletes your message)
`params`	`object`	Yes	sendMessage: space, text, threadKey · listSpaces: maxResults · listMessages: space, maxResults · updateMessage: messageName, text* · deleteMessage: messageName*

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/google-workspace-api" \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"service": "chat", "operation": "sendMessage", "params": {"space":"<space>","text":"<message>"}}'

POST

/google-workspace-api (meet)

Google Meet: createSpace, getSpace, endConference. Runs as the Google account connected in Settings → Integrations.

Key permission: integrations

Request body

Field	Type	Required	Description
`service`	`string`	Yes	"meet"
`operation`	`string`	Yes	createSpace (A new Meet link you can share or put in an event); getSpace (Link, code and whether a call is in progress); endConference (Ends the call in progress (meetings this app created))
`params`	`object`	Yes	createSpace: accessType · getSpace: space* · endConference: space*

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/google-workspace-api" \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"service": "meet", "operation": "createSpace", "params": {}}'

Google Photos

Access Google Photos albums and media items via OAuth2

POST

/google-photos-api

Google Photos API - list albums, get media items, and search photos. Requires Google OAuth connection.

Key permission: integrations

Request body

Field	Type	Required	Description
`operation`	`string`	Yes	Operation: listAlbums, getAlbum, listMediaItems, getMediaItem, searchMediaItems
`userId`	`string`	Yes	User ID (must have connected Google account with Photos scope)
`albumId`	`string`	No	Album ID (for getAlbum, listMediaItems)
`mediaItemId`	`string`	No	Media item ID (for getMediaItem)
`pageSize`	`integer`	No	Number of items to returnDefault: `25`
`pageToken`	`string`	No	Pagination token for next page
`filters`	`object`	No	Search filters (for search-media): { dateFilter, contentFilter, mediaTypeFilter }

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/google-photos-api" \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "operation": "list-albums",
    "userId": "user_abc123",
    "pageSize": 20
  }'

Jira

Jira integration for project management, issue tracking, and sprint planning

POST

/jira-integration

Jira API - list projects, create/update issues, add comments, and manage sprints.

Key permission: integrations

Request body

Field	Type	Required	Description
`operation`	`string`	Yes	Operation: list-projects, get-issue, create-issue, update-issue, search-issues, add-comment, get-transitions, transition-issue
`userId`	`string`	Yes	User ID
`projectKey`	`string`	No	Jira project key (e.g., "PROJ")
`issueKey`	`string`	No	Issue key (e.g., "PROJ-123")
`summary`	`string`	No	Issue summary (for create)
`description`	`string`	No	Issue description (for create)
`issueType`	`string`	No	Issue type: Task, Bug, Story, EpicDefault: `Task`
`transitionId`	`string`	No	Transition ID (for transition-issue)
`comment`	`string`	No	Comment body (for add-comment)
`jql`	`string`	No	JQL query string (for search-issues)
`maxResults`	`integer`	No	Max results for search-issuesDefault: `50`
`labels`	`array`	No	Array of label names (for create/update)
`assignee`	`string`	No	Jira accountId of assignee
`customFields`	`object`	No	Custom field key-value pairs (pass-through)

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/jira-integration" \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "operation": "create-issue",
    "userId": "user_abc123",
    "projectKey": "PROJ",
    "summary": "Implement OAuth refresh",
    "description": "Add automatic token refresh for Google OAuth",
    "issueType": "Task"
  }'

Facebook Ads

Facebook/Meta Ads API for campaign management, performance insights, and AI-powered optimization

POST

/facebook-ads-api

Facebook Ads API - access ad accounts, campaigns, and performance metrics

Request body

Field	Type	Required	Description
`operation`	`string`	Yes	Operation: get-ad-accounts, get-campaigns, get-adsets, get-ads, get-insights, get-account-insights, analyze-performance
`userId`	`string`	No	Optional. Defaults to the owner of the API key; any other user is rejected (403).
`params`	`object`	No	Operation-specific parameters

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/facebook-ads-api" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "operation": "get-ad-accounts",
    "userId": "user_abc123"
  }'

POST

/facebook-ads-api (get-campaigns)

Get all campaigns for an ad account

Request body

Field	Type	Required	Description
`operation`	`string`	Yes	Set to "get-campaigns"
`userId`	`string`	No	Optional. Defaults to the owner of the API key; any other user is rejected (403).
`params`	`object`	Yes	{ adAccountId: "act_123...", limit: 50 }

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/facebook-ads-api" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "operation": "get-campaigns",
    "userId": "user_abc123",
    "params": {
      "adAccountId": "act_123456789",
      "limit": 20
    }
  }'

POST

/facebook-ads-api (get-insights)

Get performance metrics and insights for campaigns, ad sets, or ads

Request body

Field	Type	Required	Description
`operation`	`string`	Yes	Set to "get-insights"
`userId`	`string`	No	Optional. Defaults to the owner of the API key; any other user is rejected (403).
`params`	`object`	Yes	{ objectId, datePreset, level, breakdown }

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/facebook-ads-api" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "operation": "get-insights",
    "userId": "user_abc123",
    "params": {
      "objectId": "act_123456789",
      "datePreset": "last_30d",
      "level": "campaign"
    }
  }'

POST

/facebook-ads-api (analyze-performance)

AI-powered analysis of ad performance with optimization recommendations

Request body

Field	Type	Required	Description
`operation`	`string`	Yes	Set to "analyze-performance"
`userId`	`string`	No	Optional. Defaults to the owner of the API key; any other user is rejected (403).
`params`	`object`	Yes	{ adAccountId, datePreset }

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/facebook-ads-api" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "operation": "analyze-performance",
    "userId": "user_abc123",
    "params": {
      "adAccountId": "act_123456789",
      "datePreset": "last_30d"
    }
  }'

Facebook Fundraisers

Create and manage Facebook fundraisers for nonprofits. Requires manage_fundraisers permission.

POST

/facebook-fundraiser-api (get-fundraisers)

Get all fundraisers for the authenticated user

Request body

Field	Type	Required	Description
`operation`	`string`	Yes	Set to "get-fundraisers"
`userId`	`string`	No	Optional. Defaults to the owner of the API key; any other user is rejected (403).
`params`	`object`	No	{ limit }

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/facebook-fundraiser-api" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "operation": "get-fundraisers",
    "userId": "user_abc123",
    "params": { "limit": 50 }
  }'

POST

/facebook-fundraiser-api (create-fundraiser)

Create a new fundraiser for a registered charity

Request body

Field	Type	Required	Description
`operation`	`string`	Yes	Set to "create-fundraiser"
`userId`	`string`	No	Optional. Defaults to the owner of the API key; any other user is rejected (403).
`params`	`object`	Yes	{ charityId, name, description, goalAmount, currency, endTime }

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/facebook-fundraiser-api" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "operation": "create-fundraiser",
    "userId": "user_abc123",
    "params": {
      "charityId": "charity_123",
      "name": "Marathon for Clean Water",
      "description": "Running to raise funds for clean water projects",
      "goalAmount": 5000,
      "currency": "USD"
    }
  }'

POST

/facebook-fundraiser-api (get-donations)

Get donations for a specific fundraiser

Request body

Field	Type	Required	Description
`operation`	`string`	Yes	Set to "get-donations"
`userId`	`string`	No	Optional. Defaults to the owner of the API key; any other user is rejected (403).
`params`	`object`	Yes	{ fundraiserId, limit }

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/facebook-fundraiser-api" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "operation": "get-donations",
    "userId": "user_abc123",
    "params": {
      "fundraiserId": "123456789",
      "limit": 100
    }
  }'

POST

/facebook-fundraiser-api (sync-amount)

Sync external fundraiser amount with Facebook fundraiser

Request body

Field	Type	Required	Description
`operation`	`string`	Yes	Set to "sync-amount"
`userId`	`string`	No	Optional. Defaults to the owner of the API key; any other user is rejected (403).
`params`	`object`	Yes	{ fundraiserId, externalAmountRaised }

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/facebook-fundraiser-api" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "operation": "sync-amount",
    "userId": "user_abc123",
    "params": {
      "fundraiserId": "123456789",
      "externalAmountRaised": 7500
    }
  }'

POST

/facebook-fundraiser-api (analyze-fundraiser)

AI-powered analysis of fundraiser performance with recommendations

Request body

Field	Type	Required	Description
`operation`	`string`	Yes	Set to "analyze-fundraiser"
`userId`	`string`	No	Optional. Defaults to the owner of the API key; any other user is rejected (403).
`params`	`object`	Yes	{ fundraiserId }

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/facebook-fundraiser-api" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "operation": "analyze-fundraiser",
    "userId": "user_abc123",
    "params": {
      "fundraiserId": "123456789"
    }
  }'

Meta Messaging

Send and receive messages via WhatsApp, Messenger, and Instagram. Requires appropriate Meta permissions.

POST

/meta-messaging-api (whatsapp-send-message)

Send a message via WhatsApp Business API

Request body

Field	Type	Required	Description
`operation`	`string`	Yes	Set to "whatsapp-send-message"
`userId`	`string`	No	Optional. Defaults to the owner of the API key; any other user is rejected (403).
`platform`	`string`	Yes	Set to "whatsapp"
`params`	`object`	Yes	{ recipientPhone, text, messageType?, mediaUrl? }

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/meta-messaging-api" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "operation": "whatsapp-send-message",
    "userId": "user_abc123",
    "platform": "whatsapp",
    "params": {
      "recipientPhone": "+1234567890",
      "text": "Hello from MultiverseAI!"
    }
  }'

POST

/meta-messaging-api (whatsapp-send-template)

Send a pre-approved WhatsApp message template

Request body

Field	Type	Required	Description
`operation`	`string`	Yes	Set to "whatsapp-send-template"
`userId`	`string`	No	Optional. Defaults to the owner of the API key; any other user is rejected (403).
`platform`	`string`	Yes	Set to "whatsapp"
`params`	`object`	Yes	{ recipientPhone, templateName, languageCode? }

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/meta-messaging-api" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "operation": "whatsapp-send-template",
    "userId": "user_abc123",
    "platform": "whatsapp",
    "params": {
      "recipientPhone": "+1234567890",
      "templateName": "order_confirmation",
      "languageCode": "en"
    }
  }'

POST

/meta-messaging-api (messenger-send-message)

Send a message via Facebook Messenger

Request body

Field	Type	Required	Description
`operation`	`string`	Yes	Set to "messenger-send-message"
`userId`	`string`	No	Optional. Defaults to the owner of the API key; any other user is rejected (403).
`platform`	`string`	Yes	Set to "messenger"
`params`	`object`	Yes	{ recipientId, text, quickReplies? }

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/meta-messaging-api" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "operation": "messenger-send-message",
    "userId": "user_abc123",
    "platform": "messenger",
    "params": {
      "recipientId": "psid_123456",
      "text": "Thanks for contacting us!",
      "quickReplies": [
        { "title": "View Order", "payload": "view_order" },
        { "title": "Contact Support", "payload": "support" }
      ]
    }
  }'

POST

/meta-messaging-api (messenger-get-conversations)

Get conversations from Facebook Messenger inbox

Request body

Field	Type	Required	Description
`operation`	`string`	Yes	Set to "messenger-get-conversations"
`userId`	`string`	No	Optional. Defaults to the owner of the API key; any other user is rejected (403).
`platform`	`string`	Yes	Set to "messenger"
`params`	`object`	No	{ limit? }

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/meta-messaging-api" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "operation": "messenger-get-conversations",
    "userId": "user_abc123",
    "platform": "messenger",
    "params": { "limit": 25 }
  }'

POST

/meta-messaging-api (instagram-send-message)

Send a direct message via Instagram

Request body

Field	Type	Required	Description
`operation`	`string`	Yes	Set to "instagram-send-message"
`userId`	`string`	No	Optional. Defaults to the owner of the API key; any other user is rejected (403).
`platform`	`string`	Yes	Set to "instagram"
`params`	`object`	Yes	{ recipientId, text, mediaUrl? }

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/meta-messaging-api" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "operation": "instagram-send-message",
    "userId": "user_abc123",
    "platform": "instagram",
    "params": {
      "recipientId": "igsid_123456",
      "text": "Thanks for your message!"
    }
  }'

POST

/meta-messaging-api (instagram-get-conversations)

Get DM conversations from Instagram inbox

Request body

Field	Type	Required	Description
`operation`	`string`	Yes	Set to "instagram-get-conversations"
`userId`	`string`	No	Optional. Defaults to the owner of the API key; any other user is rejected (403).
`platform`	`string`	Yes	Set to "instagram"
`params`	`object`	No	{ limit? }

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/meta-messaging-api" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "operation": "instagram-get-conversations",
    "userId": "user_abc123",
    "platform": "instagram",
    "params": { "limit": 25 }
  }'

TikTok Ads

TikTok Business API for ad management, campaigns, audiences, pixels, and automated rules

POST

/tiktok-api

TikTok Business API - manage advertisers, campaigns, ad groups, and ads. Requires TikTok OAuth connection.

Request body

Field	Type	Required	Description
`operation`	`string`	Yes	Operation: get-advertiser-info, get-campaigns, get-adgroups, get-ads, get-reporting, get-audiences, get-pixels
`userId`	`string`	No	Optional. Defaults to the owner of the API key; any other user is rejected (403).
`advertiserId`	`string`	No	TikTok Advertiser ID
`params`	`object`	No	Operation-specific parameters

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/tiktok-api" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "operation": "get-advertiser-info",
    "userId": "user_abc123"
  }'

POST

/tiktok-ads-api

TikTok Ads Management - create, update, and manage ad campaigns and ad groups

Request body

Field	Type	Required	Description
`operation`	`string`	Yes	Operation: create-campaign, update-campaign, create-adgroup, update-adgroup, create-ad, get-ad-insights
`userId`	`string`	No	Optional. Defaults to the owner of the API key; any other user is rejected (403).
`advertiserId`	`string`	Yes	TikTok Advertiser ID
`params`	`object`	Yes	Campaign/ad group/ad parameters

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/tiktok-ads-api" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "operation": "get-ad-insights",
    "userId": "user_abc123",
    "advertiserId": "12345678",
    "params": {
      "dateRange": "last_7_days",
      "metrics": ["impressions", "clicks", "spend", "conversions"]
    }
  }'

POST

/tiktok-pixel-api

TikTok Pixel Management - create, manage, and monitor pixels and conversion events

Request body

Field	Type	Required	Description
`operation`	`string`	Yes	Operation: create-pixel, get-pixels, get-pixel-stats, manage-events
`userId`	`string`	No	Optional. Defaults to the owner of the API key; any other user is rejected (403).
`advertiserId`	`string`	Yes	TikTok Advertiser ID
`params`	`object`	No	Pixel-specific parameters

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/tiktok-pixel-api" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "operation": "get-pixels",
    "userId": "user_abc123",
    "advertiserId": "12345678"
  }'

POST

/tiktok-automated-rules-api

TikTok Automated Rules - create and manage automated optimization rules for ads

Request body

Field	Type	Required	Description
`operation`	`string`	Yes	Operation: create-rule, get-rules, update-rule, delete-rule, get-rule-results
`userId`	`string`	No	Optional. Defaults to the owner of the API key; any other user is rejected (403).
`advertiserId`	`string`	Yes	TikTok Advertiser ID
`params`	`object`	No	Rule-specific parameters

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/tiktok-automated-rules-api" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "operation": "get-rules",
    "userId": "user_abc123",
    "advertiserId": "12345678"
  }'

POST

/tiktok-rf-api

TikTok Reach & Frequency - plan and manage R&F campaigns with guaranteed delivery

Request body

Field	Type	Required	Description
`operation`	`string`	Yes	Operation: estimate-inventory, create-order, get-orders
`userId`	`string`	No	Optional. Defaults to the owner of the API key; any other user is rejected (403).
`advertiserId`	`string`	Yes	TikTok Advertiser ID
`params`	`object`	No	R&F campaign parameters

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/tiktok-rf-api" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "operation": "estimate-inventory",
    "userId": "user_abc123",
    "advertiserId": "12345678",
    "params": {
      "targeting": { "location": ["US"], "age": ["18-24", "25-34"] },
      "budget": 5000
    }
  }'

Email Templates

Styled email templates for transactional emails (welcome, password reset, notifications)

POST

/email-templates

Generate and send styled HTML emails using pre-built templates

Request body

Field	Type	Required	Description
`template`	`string`	Yes	Template: welcome, password_reset, notification, team_invite, task_assigned
`to`	`string`	Yes	Recipient email address
`data`	`object`	Yes	Template-specific data (e.g., { userName, teamName, inviteLink })

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/email-templates" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "template": "welcome",
    "to": "newuser@example.com",
    "data": {
      "userName": "Jane Doe",
      "loginUrl": "https://multiverse-ai.lovable.app/auth"
    }
  }'

Task Notifications

Send task-related notifications for due dates, assignments, and @mentions

POST

/task-notifications

Trigger task notifications for due date reminders, assignment alerts, and comment mentions

Key permission: tasks

Request body

Field	Type	Required	Description
`type`	`string`	Yes	Notification type: due-reminder, assignment, mention
`taskId`	`string`	Yes	Task ID
`userId`	`string`	Yes	User ID triggering the notification
`targetUserId`	`string`	No	Target user to notify (for assignment/mention)
`message`	`string`	No	Optional custom message

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/task-notifications" \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "type": "assignment",
    "taskId": "task_abc123",
    "userId": "user_sender",
    "targetUserId": "user_recipient",
    "message": "You have been assigned to: Fix login bug"
  }'

MCP Server

Model Context Protocol (MCP) server for AI agent integrations. Allows external AI agents like Claude, Cursor, Windsurf, or custom agents to use MultiverseAI programmatically via the Streamable HTTP transport (JSON-RPC over POST): agents, workflows, contexts, the prompt queue, chat, images, video, agent tools, tasks, teams and social posting. Tools that call AI models use your credits; developer-key switches (tasks, workflows, agents) apply as in the REST API.

POST

/mcp-server

MCP Streamable HTTP endpoint. Authenticate with your API key and call tools/list to discover all 51 available tools. Tool errors (including MultiverseAI API errors such as INSUFFICIENT_CREDITS) come back as results with isError: true and the error message. Compatible with any MCP client (Claude Desktop, Cursor, Windsurf, custom agents).

Request body

Field	Type	Required	Description
`jsonrpc`	`string`	Yes	Must be "2.0"Default: `2.0`
`id`	`number`	Yes	Request ID
`method`	`string`	Yes	MCP method: tools/list or tools/call

curl -X POST "https://ralcxfiefghnpadaddzu.supabase.co/functions/v1/mcp-server" \
  -H "Authorization: Bearer mverse_k_YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "jsonrpc": "2.0",
    "id": 1,
    "method": "tools/list"
  }'

POST

/mcp-server

list_agents — List all AI agents. Filter by status (draft/published/archived) and pagination.

Key permission: agents

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `list_agents`
`params.arguments.status`	`string`	No	Filter: draft, published, archived
`params.arguments.include_archived`	`boolean`	No	Include archived agents (default false)
`params.arguments.limit`	`number`	No	Max results (default 50, max 200)

{
  "jsonrpc": "2.0",
  "id": 2,
  "method": "tools/call",
  "params": {
    "name": "list_agents",
    "arguments": {
      "status": "published",
      "limit": 10
    }
  }
}

POST

/mcp-server

get_agent — Get detailed agent info including linked context, workflows, custom tools, and version history.

Key permission: agents

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `get_agent`
`params.arguments.agent_id`	`string`	Yes	The agent UUID

{
  "jsonrpc": "2.0",
  "id": 3,
  "method": "tools/call",
  "params": {
    "name": "get_agent",
    "arguments": {
      "agent_id": "abc-123"
    }
  }
}

POST

/mcp-server

create_agent — Create a new AI agent with name, instructions, model, personality, and optional context attachment.

Key permission: agents

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `create_agent`
`params.arguments.name`	`string`	Yes	Agent name
`params.arguments.description`	`string`	No	Agent description
`params.arguments.instructions`	`string`	No	System instructions
`params.arguments.default_model`	`string`	No	AI model (default: google/gemini-3-flash-preview)
`params.arguments.personality_template`	`string`	No	professional \| casual \| technical \| friendly \| creative
`params.arguments.response_length`	`string`	No	concise \| balanced \| detailed
`params.arguments.context_id`	`string`	No	UUID of a prompt context to attach
`params.arguments.status`	`string`	No	draft \| published (default: draft)

{
  "jsonrpc": "2.0",
  "id": 4,
  "method": "tools/call",
  "params": {
    "name": "create_agent",
    "arguments": {
      "name": "Sales Assistant",
      "description": "Helps qualify leads and answer product questions",
      "instructions": "You are a knowledgeable sales assistant for MultiverseAI...",
      "default_model": "google/gemini-3-flash-preview",
      "personality_template": "professional",
      "response_length": "balanced",
      "status": "draft"
    }
  }
}

POST

/mcp-server

update_agent — Update an agent's name, instructions, model, personality, status, or linked context.

Key permission: agents

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `update_agent`
`params.arguments.agent_id`	`string`	Yes
`params.arguments.name`	`string`	No
`params.arguments.instructions`	`string`	No
`params.arguments.default_model`	`string`	No
`params.arguments.personality_template`	`string`	No
`params.arguments.response_length`	`string`	No
`params.arguments.context_id`	`string`	No
`params.arguments.status`	`string`	No	draft \| published \| archived

{
  "jsonrpc": "2.0",
  "id": 5,
  "method": "tools/call",
  "params": {
    "name": "update_agent",
    "arguments": {
      "agent_id": "abc-123",
      "status": "published",
      "instructions": "Updated instructions..."
    }
  }
}

POST

/mcp-server

delete_agent — Permanently delete an AI agent. Irreversible.

Key permission: agents

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `delete_agent`
`params.arguments.agent_id`	`string`	Yes

{
  "jsonrpc": "2.0",
  "id": 6,
  "method": "tools/call",
  "params": {
    "name": "delete_agent",
    "arguments": {
      "agent_id": "abc-123"
    }
  }
}

POST

/mcp-server

list_workflows — List all workflows. Filter by active status, with pagination.

Key permission: workflows

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `list_workflows`
`params.arguments.limit`	`number`	No	Max results (default 50, max 200)
`params.arguments.is_active`	`boolean`	No	Filter by active status

{
  "jsonrpc": "2.0",
  "id": 7,
  "method": "tools/call",
  "params": {
    "name": "list_workflows",
    "arguments": {
      "is_active": true,
      "limit": 10
    }
  }
}

POST

/mcp-server

get_workflow — Get full workflow details including all nodes and edges.

Key permission: workflows

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `get_workflow`
`params.arguments.workflow_id`	`string`	Yes

{
  "jsonrpc": "2.0",
  "id": 8,
  "method": "tools/call",
  "params": {
    "name": "get_workflow",
    "arguments": {
      "workflow_id": "wf-123"
    }
  }
}

POST

/mcp-server

create_workflow — Create a new workflow with optional initial nodes, edges, category, and tags.

Key permission: workflows

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `create_workflow`
`params.arguments.name`	`string`	Yes
`params.arguments.description`	`string`	No
`params.arguments.nodes`	`array`	No	Initial node definitions (JSON)
`params.arguments.edges`	`array`	No	Initial edge definitions (JSON)
`params.arguments.category`	`string`	No
`params.arguments.tags`	`array`	No
`params.arguments.is_active`	`boolean`	No

{
  "jsonrpc": "2.0",
  "id": 9,
  "method": "tools/call",
  "params": {
    "name": "create_workflow",
    "arguments": {
      "name": "Email Digest",
      "description": "Daily summary workflow",
      "is_active": true
    }
  }
}

POST

/mcp-server

update_workflow — Update a workflow's name, description, nodes, edges, status, category, tags, or active status.

Key permission: workflows

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `update_workflow`
`params.arguments.workflow_id`	`string`	Yes
`params.arguments.name`	`string`	No
`params.arguments.description`	`string`	No
`params.arguments.nodes`	`array`	No
`params.arguments.edges`	`array`	No
`params.arguments.status`	`string`	No
`params.arguments.is_active`	`boolean`	No

{
  "jsonrpc": "2.0",
  "id": 10,
  "method": "tools/call",
  "params": {
    "name": "update_workflow",
    "arguments": {
      "workflow_id": "wf-123",
      "is_active": false
    }
  }
}

POST

/mcp-server

delete_workflow — Permanently delete a workflow.

Key permission: workflows

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `delete_workflow`
`params.arguments.workflow_id`	`string`	Yes

{
  "jsonrpc": "2.0",
  "id": 11,
  "method": "tools/call",
  "params": {
    "name": "delete_workflow",
    "arguments": {
      "workflow_id": "wf-123"
    }
  }
}

POST

/mcp-server

list_workflow_executions — List recent execution history for a workflow. Filter by status (e.g. running/completed/failed).

Key permission: workflows

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `list_workflow_executions`
`params.arguments.workflow_id`	`string`	Yes
`params.arguments.limit`	`number`	No	Max results (default 20)
`params.arguments.status`	`string`	No	Filter, e.g. running, completed, failed

{
  "jsonrpc": "2.0",
  "id": 12,
  "method": "tools/call",
  "params": {
    "name": "list_workflow_executions",
    "arguments": {
      "workflow_id": "wf-123",
      "status": "completed",
      "limit": 5
    }
  }
}

POST

/mcp-server

list_contexts — List all prompt contexts (knowledge bases). Optionally include archived.

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `list_contexts`
`params.arguments.limit`	`number`	No
`params.arguments.include_archived`	`boolean`	No

{
  "jsonrpc": "2.0",
  "id": 13,
  "method": "tools/call",
  "params": {
    "name": "list_contexts",
    "arguments": {
      "limit": 20
    }
  }
}

POST

/mcp-server

get_context — Get a context with its conversation messages and knowledge entries (text, Q&A pairs, URL sources).

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `get_context`
`params.arguments.context_id`	`string`	Yes
`params.arguments.include_messages`	`boolean`	No	Include conversation messages (default true)
`params.arguments.include_knowledge`	`boolean`	No	Include knowledge entries (default true)

{
  "jsonrpc": "2.0",
  "id": 14,
  "method": "tools/call",
  "params": {
    "name": "get_context",
    "arguments": {
      "context_id": "ctx-456",
      "include_knowledge": true,
      "include_messages": false
    }
  }
}

POST

/mcp-server

create_context — Create a new prompt context (knowledge base) with system prompt, model, and tags.

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `create_context`
`params.arguments.name`	`string`	Yes
`params.arguments.description`	`string`	No
`params.arguments.system_prompt`	`string`	No
`params.arguments.model`	`string`	No
`params.arguments.temperature`	`number`	No
`params.arguments.tags`	`array`	No	Array of tag strings

{
  "jsonrpc": "2.0",
  "id": 15,
  "method": "tools/call",
  "params": {
    "name": "create_context",
    "arguments": {
      "name": "API Docs KB",
      "description": "Knowledge base for API documentation",
      "system_prompt": "You are a documentation expert.",
      "tags": [
        "docs",
        "api"
      ]
    }
  }
}

POST

/mcp-server

add_knowledge_entry — Add a text, Q&A pair, URL source, document chunk, or CSV row to a context's knowledge base.

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `add_knowledge_entry`
`params.arguments.context_id`	`string`	Yes
`params.arguments.content`	`string`	Yes
`params.arguments.entry_type`	`string`	No	text \| qa_pair \| url_source \| document_chunk \| csv_row (default text)
`params.arguments.title`	`string`	No
`params.arguments.answer`	`string`	No	Answer for qa_pair entries
`params.arguments.source_url`	`string`	No
`params.arguments.tags`	`array`	No
`params.arguments.priority`	`number`	No	Higher = more important

{
  "jsonrpc": "2.0",
  "id": 16,
  "method": "tools/call",
  "params": {
    "name": "add_knowledge_entry",
    "arguments": {
      "context_id": "ctx-456",
      "title": "Return Policy",
      "content": "What is your return policy?",
      "entry_type": "qa_pair",
      "answer": "We offer 30-day returns on all products.",
      "priority": 5
    }
  }
}

POST

/mcp-server

add_context_message — Add a message (user/assistant/system) to a context's conversation history.

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `add_context_message`
`params.arguments.context_id`	`string`	Yes
`params.arguments.role`	`string`	Yes	user \| assistant \| system
`params.arguments.content`	`string`	Yes
`params.arguments.model`	`string`	No

{
  "jsonrpc": "2.0",
  "id": 17,
  "method": "tools/call",
  "params": {
    "name": "add_context_message",
    "arguments": {
      "context_id": "ctx-456",
      "role": "user",
      "content": "What models do you support?"
    }
  }
}

POST

/mcp-server

list_prompts — List prompt queue items. Filter by status (pending/processing/completed/failed) or model.

Key permission: workflows

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `list_prompts`
`params.arguments.status`	`string`	No	Filter: pending, processing, completed, failed
`params.arguments.model`	`string`	No	Filter by model
`params.arguments.limit`	`number`	No
`params.arguments.offset`	`number`	No

{
  "jsonrpc": "2.0",
  "id": 18,
  "method": "tools/call",
  "params": {
    "name": "list_prompts",
    "arguments": {
      "status": "completed",
      "limit": 5
    }
  }
}

POST

/mcp-server

get_prompt — Get full prompt details including the AI response, tokens used, and execution metadata.

Key permission: workflows

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `get_prompt`
`params.arguments.prompt_id`	`string`	Yes

{
  "jsonrpc": "2.0",
  "id": 19,
  "method": "tools/call",
  "params": {
    "name": "get_prompt",
    "arguments": {
      "prompt_id": "p-1"
    }
  }
}

POST

/mcp-server

submit_prompt — Add a prompt to the queue without running it (enqueue_prompt runs or schedules it).

Key permission: workflows

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `submit_prompt`
`params.arguments.prompt`	`string`	Yes	The prompt text
`params.arguments.model`	`string`	No	Model id (see list_models)
`params.arguments.destination`	`string`	No	Output: file \| chat \| email
`params.arguments.priority`	`string`	No	critical \| high \| normal \| low

{
  "jsonrpc": "2.0",
  "id": 20,
  "method": "tools/call",
  "params": {
    "name": "submit_prompt",
    "arguments": {
      "prompt": "Summarize today's tasks",
      "priority": "high"
    }
  }
}

POST

/mcp-server

list_conversations — List AI chat conversations with message counts. Optionally include archived.

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `list_conversations`
`params.arguments.limit`	`number`	No	Max results (default 30)
`params.arguments.include_archived`	`boolean`	No

{
  "jsonrpc": "2.0",
  "id": 21,
  "method": "tools/call",
  "params": {
    "name": "list_conversations",
    "arguments": {
      "limit": 10
    }
  }
}

POST

/mcp-server

get_conversation_messages — Get all messages in a conversation (up to 200). Verifies conversation ownership.

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `get_conversation_messages`
`params.arguments.conversation_id`	`string`	Yes
`params.arguments.limit`	`number`	No	Max messages (default 200)

{
  "jsonrpc": "2.0",
  "id": 22,
  "method": "tools/call",
  "params": {
    "name": "get_conversation_messages",
    "arguments": {
      "conversation_id": "conv-1",
      "limit": 50
    }
  }
}

POST

/mcp-server

get_usage_stats — Get AI usage statistics: total tokens, costs, request counts, success rate, avg response time, breakdown by provider.

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `get_usage_stats`
`params.arguments.days`	`number`	No	Lookback days (default 30, max 365)
`params.arguments.provider`	`string`	No	Filter by provider

{
  "jsonrpc": "2.0",
  "id": 23,
  "method": "tools/call",
  "params": {
    "name": "get_usage_stats",
    "arguments": {
      "days": 7
    }
  }
}

POST

/mcp-server

get_agent_analytics — Get event-level analytics for a specific agent: event counts by type and recent activity.

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `get_agent_analytics`
`params.arguments.agent_id`	`string`	Yes
`params.arguments.days`	`number`	No	Lookback days (default 30, max 365)

{
  "jsonrpc": "2.0",
  "id": 24,
  "method": "tools/call",
  "params": {
    "name": "get_agent_analytics",
    "arguments": {
      "agent_id": "abc-123",
      "days": 14
    }
  }
}

POST

/mcp-server

list_tasks — List tasks from task boards. Filter by status (backlog/todo/in-progress/review/done), priority (low/medium/high/urgent), board, or assignee.

Key permission: tasks

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `list_tasks`
`params.arguments.status`	`string`	No	Filter: backlog, todo, in-progress, review, done
`params.arguments.priority`	`string`	No	Filter: low, medium, high, urgent
`params.arguments.board_id`	`string`	No
`params.arguments.limit`	`number`	No

{
  "jsonrpc": "2.0",
  "id": 25,
  "method": "tools/call",
  "params": {
    "name": "list_tasks",
    "arguments": {
      "status": "in-progress",
      "priority": "high"
    }
  }
}

POST

/mcp-server

create_task — Create a new task with title, description, status, priority, tags, due date, and assignee.

Key permission: tasks

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `create_task`
`params.arguments.title`	`string`	Yes
`params.arguments.description`	`string`	No
`params.arguments.status`	`string`	No	backlog \| todo \| in-progress \| review \| done (default: todo)
`params.arguments.priority`	`string`	No	low \| medium \| high \| urgent (default: medium)
`params.arguments.board_id`	`string`	No	Defaults to your default board
`params.arguments.tags`	`array`	No
`params.arguments.due_date`	`string`	No	ISO date string
`params.arguments.assignee`	`string`	No

{
  "jsonrpc": "2.0",
  "id": 26,
  "method": "tools/call",
  "params": {
    "name": "create_task",
    "arguments": {
      "title": "Add OAuth refresh",
      "description": "Implement automatic token refresh for Google OAuth",
      "status": "todo",
      "priority": "high",
      "tags": [
        "feature",
        "auth"
      ],
      "due_date": "2026-04-25"
    }
  }
}

POST

/mcp-server

update_task — Update a task's title, description, status, priority, tags, due date, or assignee.

Key permission: tasks

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `update_task`
`params.arguments.task_id`	`string`	Yes
`params.arguments.title`	`string`	No
`params.arguments.description`	`string`	No
`params.arguments.status`	`string`	No
`params.arguments.priority`	`string`	No
`params.arguments.tags`	`array`	No
`params.arguments.due_date`	`string`	No
`params.arguments.assignee`	`string`	No

{
  "jsonrpc": "2.0",
  "id": 27,
  "method": "tools/call",
  "params": {
    "name": "update_task",
    "arguments": {
      "task_id": "task-1",
      "status": "done"
    }
  }
}

POST

/mcp-server

list_notifications — List user notifications. Filter by read status or type (success/error/info).

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `list_notifications`
`params.arguments.unread_only`	`boolean`	No
`params.arguments.type`	`string`	No	Filter: success, error, info
`params.arguments.limit`	`number`	No	Max results (default 30)

{
  "jsonrpc": "2.0",
  "id": 28,
  "method": "tools/call",
  "params": {
    "name": "list_notifications",
    "arguments": {
      "unread_only": true,
      "limit": 10
    }
  }
}

POST

/mcp-server

list_integrations — List all configured integrations (email, Slack, Google Drive, webhooks) with their enabled status.

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `list_integrations`

{
  "jsonrpc": "2.0",
  "id": 29,
  "method": "tools/call",
  "params": {
    "name": "list_integrations",
    "arguments": {}
  }
}

POST

/mcp-server

list_models — List the available AI models (id, provider, modality, tier, tool support, cost).

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `list_models`
`params.arguments.modality`	`string`	No	text \| image \| video

{
  "jsonrpc": "2.0",
  "id": 100,
  "method": "tools/call",
  "params": {
    "name": "list_models",
    "arguments": {
      "modality": "text"
    }
  }
}

POST

/mcp-server

chat — One-shot chat completion with any model. Pass messages (OpenAI format) or prompt. Uses credits.

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `chat`
`params.arguments.prompt`	`string`	No	Shortcut for a single user message
`params.arguments.system`	`string`	No	System prompt (with prompt)
`params.arguments.messages`	`array`	No	[{ role, content }] — content may include image_url parts
`params.arguments.model`	`string`	No	Model id (default openrouter/openai/gpt-6-luna)
`params.arguments.tools`	`boolean \| string[]`	No	true = all MultiverseAI agent tools, or tool names from list_agent_tools
`params.arguments.max_tool_iterations`	`number`	No	1–20
`params.arguments.temperature`	`number`	No
`params.arguments.max_tokens`	`number`	No

{
  "jsonrpc": "2.0",
  "id": 101,
  "method": "tools/call",
  "params": {
    "name": "chat",
    "arguments": {
      "prompt": "Summarize the latest news about AI agents",
      "tools": [
        "web_search"
      ]
    }
  }
}

POST

/mcp-server

generate_image — Generate images from a prompt, optionally working from reference images; returns image URLs. Uses credits.

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `generate_image`
`params.arguments.prompt`	`string`	Yes
`params.arguments.model`	`string`	No	Image model id (default gpt-image-1.5)
`params.arguments.n`	`number`	No	1–4
`params.arguments.image`	`string[]`	No	Reference images (https or data: URLs, PNG/JPEG/WebP, up to 4)

{
  "jsonrpc": "2.0",
  "id": 102,
  "method": "tools/call",
  "params": {
    "name": "generate_image",
    "arguments": {
      "prompt": "A watercolor fox in a misty forest"
    }
  }
}

POST

/mcp-server

create_video — Start a video generation job; poll get_video until completed. Uses credits.

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `create_video`
`params.arguments.prompt`	`string`	Yes
`params.arguments.model`	`string`	No	Video model id (list_models modality=video)
`params.arguments.duration`	`number`	No
`params.arguments.aspect_ratio`	`string`	No	e.g. 16:9, 9:16
`params.arguments.image_url`	`string`	No	Start image (https:// or data:image/…)

{
  "jsonrpc": "2.0",
  "id": 103,
  "method": "tools/call",
  "params": {
    "name": "create_video",
    "arguments": {
      "prompt": "Drone shot over a coastline at sunrise",
      "aspect_ratio": "16:9"
    }
  }
}

POST

/mcp-server

get_video — Get a video job: status (queued/running/saving/completed/failed/cancelled) and video_url.

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `get_video`
`params.arguments.video_id`	`string`	Yes

{
  "jsonrpc": "2.0",
  "id": 104,
  "method": "tools/call",
  "params": {
    "name": "get_video",
    "arguments": {
      "video_id": "VIDEO_ID"
    }
  }
}

POST

/mcp-server

list_agent_tools — List the MultiverseAI agent tools. Pass name for one tool with its parameter schema.

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `list_agent_tools`
`params.arguments.name`	`string`	No
`params.arguments.include_parameters`	`boolean`	No	Include every parameter schema (large)

{
  "jsonrpc": "2.0",
  "id": 105,
  "method": "tools/call",
  "params": {
    "name": "list_agent_tools",
    "arguments": {}
  }
}

POST

/mcp-server

run_agent_tool — Run one agent tool directly. Tools with requires_approval need confirm: true.

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `run_agent_tool`
`params.arguments.name`	`string`	Yes
`params.arguments.arguments`	`object`	No
`params.arguments.confirm`	`boolean`	No

{
  "jsonrpc": "2.0",
  "id": 106,
  "method": "tools/call",
  "params": {
    "name": "run_agent_tool",
    "arguments": {
      "name": "web_search",
      "arguments": {
        "query": "MultiverseAI"
      }
    }
  }
}

POST

/mcp-server

link_previews — Fetch title, description, image and site name for URLs.

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `link_previews`
`params.arguments.urls`	`array`	Yes

{
  "jsonrpc": "2.0",
  "id": 107,
  "method": "tools/call",
  "params": {
    "name": "link_previews",
    "arguments": {
      "urls": [
        "https://multiverseai.com"
      ]
    }
  }
}

POST

/mcp-server

get_credits — Get your MultiverseAI credit balance.

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `get_credits`
`params.arguments.fresh`	`boolean`	No	Skip the 60-second cache

{
  "jsonrpc": "2.0",
  "id": 108,
  "method": "tools/call",
  "params": {
    "name": "get_credits",
    "arguments": {}
  }
}

POST

/mcp-server

chat_with_agent — Send a message to one of your agents. Pass conversation_id from a reply to continue it.

Key permission: agents

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `chat_with_agent`
`params.arguments.agent_id`	`string`	Yes
`params.arguments.message`	`string`	Yes
`params.arguments.conversation_id`	`string`	No

{
  "jsonrpc": "2.0",
  "id": 109,
  "method": "tools/call",
  "params": {
    "name": "chat_with_agent",
    "arguments": {
      "agent_id": "AGENT_ID",
      "message": "What can you help me with?"
    }
  }
}

POST

/mcp-server

run_workflow — Run a workflow on the server. Returns the execution id, or the result with wait: true.

Key permission: workflows

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `run_workflow`
`params.arguments.workflow_id`	`string`	Yes
`params.arguments.inputs`	`object`	No
`params.arguments.wait`	`boolean`	No	Default false

{
  "jsonrpc": "2.0",
  "id": 110,
  "method": "tools/call",
  "params": {
    "name": "run_workflow",
    "arguments": {
      "workflow_id": "WORKFLOW_ID",
      "inputs": {
        "topic": "AI"
      }
    }
  }
}

POST

/mcp-server

get_execution — Get one workflow run: status, node outputs, execution data, error.

Key permission: workflows

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `get_execution`
`params.arguments.execution_id`	`string`	Yes

{
  "jsonrpc": "2.0",
  "id": 111,
  "method": "tools/call",
  "params": {
    "name": "get_execution",
    "arguments": {
      "execution_id": "EXECUTION_ID"
    }
  }
}

POST

/mcp-server

create_conversation — Start a new chat conversation.

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `create_conversation`
`params.arguments.title`	`string`	No
`params.arguments.model`	`string`	No
`params.arguments.folder_id`	`string`	No

{
  "jsonrpc": "2.0",
  "id": 112,
  "method": "tools/call",
  "params": {
    "name": "create_conversation",
    "arguments": {
      "title": "Research notes"
    }
  }
}

POST

/mcp-server

send_conversation_message — Send a message in a conversation and get the reply; both are saved. Uses credits.

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `send_conversation_message`
`params.arguments.conversation_id`	`string`	Yes
`params.arguments.content`	`string`	Yes
`params.arguments.model`	`string`	No
`params.arguments.tools`	`boolean \| string[]`	No	true = all MultiverseAI agent tools, or tool names from list_agent_tools
`params.arguments.system_prompt`	`string`	No
`params.arguments.temperature`	`number`	No

{
  "jsonrpc": "2.0",
  "id": 113,
  "method": "tools/call",
  "params": {
    "name": "send_conversation_message",
    "arguments": {
      "conversation_id": "CONVERSATION_ID",
      "content": "Continue from where we left off"
    }
  }
}

POST

/mcp-server

list_queue — List prompt queue items, newest first.

Key permission: workflows

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `list_queue`
`params.arguments.status`	`string`	No	Comma-separated, e.g. pending,running
`params.arguments.limit`	`number`	No
`params.arguments.offset`	`number`	No

{
  "jsonrpc": "2.0",
  "id": 114,
  "method": "tools/call",
  "params": {
    "name": "list_queue",
    "arguments": {
      "status": "pending,running"
    }
  }
}

POST

/mcp-server

enqueue_prompt — Queue a prompt (or a workflow run) and run it now, or schedule it. Uses credits.

Key permission: workflows

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `enqueue_prompt`
`params.arguments.prompt`	`string`	No	Required unless workflow_id is given
`params.arguments.workflow_id`	`string`	No
`params.arguments.model`	`string`	No
`params.arguments.priority`	`string`	No	critical \| high \| normal \| low
`params.arguments.destination`	`string`	No
`params.arguments.scheduled_for`	`string`	No	ISO 8601, in the future
`params.arguments.run`	`boolean`	No	Default true
`params.arguments.wait`	`boolean`	No

{
  "jsonrpc": "2.0",
  "id": 115,
  "method": "tools/call",
  "params": {
    "name": "enqueue_prompt",
    "arguments": {
      "prompt": "Draft a weekly status update",
      "priority": "high"
    }
  }
}

POST

/mcp-server

cancel_queue_item — Cancel a pending, scheduled or retrying queue item.

Key permission: workflows

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `cancel_queue_item`
`params.arguments.queue_id`	`string`	Yes

{
  "jsonrpc": "2.0",
  "id": 116,
  "method": "tools/call",
  "params": {
    "name": "cancel_queue_item",
    "arguments": {
      "queue_id": "QUEUE_ID"
    }
  }
}

POST

/mcp-server

delete_task — Delete a task permanently.

Key permission: tasks

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `delete_task`
`params.arguments.task_id`	`string`	Yes

{
  "jsonrpc": "2.0",
  "id": 117,
  "method": "tools/call",
  "params": {
    "name": "delete_task",
    "arguments": {
      "task_id": "TASK_ID"
    }
  }
}

POST

/mcp-server

list_boards — List your task boards.

Key permission: tasks

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `list_boards`
`params.arguments.limit`	`number`	No
`params.arguments.offset`	`number`	No

{
  "jsonrpc": "2.0",
  "id": 118,
  "method": "tools/call",
  "params": {
    "name": "list_boards",
    "arguments": {}
  }
}

POST

/mcp-server

list_teams — List your teams; pass team_id for one team with its members.

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `list_teams`
`params.arguments.team_id`	`string`	No

{
  "jsonrpc": "2.0",
  "id": 119,
  "method": "tools/call",
  "params": {
    "name": "list_teams",
    "arguments": {}
  }
}

POST

/mcp-server

list_platforms — Show which posting platforms (twitter, tiktok, instagram, facebook) are connected.

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `list_platforms`

{
  "jsonrpc": "2.0",
  "id": 120,
  "method": "tools/call",
  "params": {
    "name": "list_platforms",
    "arguments": {}
  }
}

POST

/mcp-server

publish_post — Publish a post now as your connected accounts.

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `publish_post`
`params.arguments.platforms`	`array`	Yes	twitter \| tiktok \| instagram \| facebook
`params.arguments.content`	`string`	No	Post text
`params.arguments.media_urls`	`array`	No	https:// media URLs (Instagram and TikTok need one)
`params.arguments.platform_options`	`object`	No	{ facebook: { page_id, link }, tiktok: { privacy_level } }

{
  "jsonrpc": "2.0",
  "id": 121,
  "method": "tools/call",
  "params": {
    "name": "publish_post",
    "arguments": {
      "platforms": [
        "twitter"
      ],
      "content": "Shipping something new today"
    }
  }
}

POST

/mcp-server

schedule_post — Schedule a post for later.

Request body

Field	Type	Required	Description
`params.name`	`string`	Yes	Default: `schedule_post`
`params.arguments.platforms`	`array`	Yes	twitter \| tiktok \| instagram \| facebook
`params.arguments.content`	`string`	No	Post text
`params.arguments.media_urls`	`array`	No	https:// media URLs (Instagram and TikTok need one)
`params.arguments.platform_options`	`object`	No	{ facebook: { page_id, link }, tiktok: { privacy_level } }
`params.arguments.scheduled_at`	`string`	Yes	ISO 8601, in the future

{
  "jsonrpc": "2.0",
  "id": 122,
  "method": "tools/call",
  "params": {
    "name": "schedule_post",
    "arguments": {
      "platforms": [
        "facebook"
      ],
      "content": "See you at the launch!",
      "scheduled_at": "2026-12-01T15:00:00Z"
    }
  }
}

Error Handling

Every /v1 error has the same shape: an error object with a message, a machine-readable code and the HTTP status. Some errors add fields such as retry_after_ms or allowed.

Error response format

{
  "error": {
    "message": "Too many requests for this API key",
    "code": "RATE_LIMITED",
    "status": 429,
    "retry_after_ms": 41250
  }
}

HTTP status codes

200

OKRequest succeeded

201

CreatedResource created (or a conversation message exchanged)

202

AcceptedStarted in the background (workflow runs)

400

Bad RequestInvalid body or parameters

401

UnauthorizedMissing, invalid, revoked or expired key or token

402

Payment RequiredNot enough MultiverseAI credits

403

ForbiddenThe key's permission switch for this endpoint is off

404

Not FoundUnknown endpoint, model, tool, or a resource you don't own

405

Method Not AllowedHTTP method not supported (allowed lists the ones that are)

409

ConflictA tool needs confirm: true

413

Payload Too LargeAn attachment is over the size limit (images: 20MB)

429

Too Many RequestsRate limit hit; wait retry_after_ms

500

Server ErrorUnexpected error

502

Bad GatewayThe model provider or an upstream service failed

Error codes

The most common values of error.code

Code	Description
`AUTH_REQUIRED`	No Authorization: Bearer header
`INVALID_API_KEY`	Developer key is invalid, revoked or expired
`INVALID_TOKEN`	User access token is invalid or expired
`FORBIDDEN`	This key isn't allowed to use the endpoint (permission switch off)
`NOT_FOUND`	Unknown endpoint, or the resource does not exist / is not yours
`MODEL_NOT_FOUND`	Unknown model id — see GET /v1/models
`INVALID_REQUEST`	Missing or malformed field
`INVALID_VALUE`	A field has the wrong type or an unsupported value
`UNSUPPORTED_TOOLS`	OpenAI function definitions were passed; use tools: true or tool names
`METHOD_NOT_ALLOWED`	HTTP method not supported for this endpoint
`RATE_LIMITED`	Too many requests; includes retry_after_ms
`INSUFFICIENT_CREDITS`	Not enough credits for an AI call
`CONFIRMATION_REQUIRED`	The tool changes data or sends messages; resend with confirm: true
`UNKNOWN_TOOL`	No tool with that name — see GET /v1/tools
`PAYLOAD_TOO_LARGE`	An attachment is over the size limit (images: 20MB)
`UPSTREAM_ERROR`	The model provider or an internal service failed, or returned an empty reply
`INTERNAL_ERROR`	Unexpected server error

Get started

Ready to get started?

Generate your API key and start building with MultiverseAI