MGPT Platform

Base URL

# Public gateway
https://mgpt.continualmi.com/api/mgpt

# Local development (backend-api)
http://127.0.0.1:8090/api/mgpt

All examples on this page use the public gateway. The local backend accepts the same bodies under the /api/mgpt prefix.

Authentication

All endpoints require a bearer token issued from the MGPT API console. Keys have the shape cmi_live_<public>.<secret> and are stored hashed in core.api_keys.

Authorization: Bearer cmi_live_<public>.<secret>

Headers

Header	Required	Notes
`Authorization`	yes	`Bearer <api-key>`
`Content-Type`	yes (POST)	`application/json`
`X-Request-Id`	no	Optional caller correlation id for logs and diagnostics.

POST/api/mgpt/chat/completions

Create chat completion

OpenAI-compatible text generation through the MGPT model gateway. MGPT resolves the public model id, routes the request to the configured execution backend, records diagnostics, and returns a chat completion response or SSE stream.

Required scope: mgpt.responses:create.

Request body

Field	Type	Required	Notes
`model`	string	yes	Model id from the Models page, for example `deepseek-v4-pro` or `deepseek-v4-flash`.
`messages`	array	yes	OpenAI-compatible chat messages.
`stream`	boolean	no	When true, returns server-sent events.
`tools`	array	no	Function-calling tool definitions forwarded to compatible providers.
`tool_choice`	string \| object	no	OpenAI-compatible tool selection.
`stream_options`	object	no	Set `{ "include_usage": true }` to include usage in streamed responses when supported.

Response body

{
  "id": "chatcmpl_...",
  "object": "chat.completion",
  "created": 1770280000,
  "model": "deepseek-v4-pro",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "Hello from MGPT."
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 12,
    "completion_tokens": 5,
    "total_tokens": 17
  }
}

Examples

curl https://mgpt.continualmi.com/api/mgpt/chat/completions \
  -H "Authorization: Bearer $CONTINUAL_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-v4-pro",
    "messages": [
      { "role": "user", "content": "Say hello in one sentence." }
    ]
  }'

POST/api/mgpt/images

Generate image

Image generation endpoint for platform and MDL media workflows. MGPT assembles the provider image message from the system prompt, description, optional references, and image configuration.

Required scope: mgpt.images:create.

Field	Type	Required	Notes
`systemPrompt`	string	yes	Image-generation system prompt.
`description`	string	yes	Main scene or asset description.
`additionalPrompt`	string	no	Short addendum appended last.
`referenceImages`	array	no	Interleaved text/image parts for references.
`imageConfig`	object	no	`imageSize`: `1K` \| `2K` \| `4K`; `aspectRatio` defaults to `16:9`.
`model`	string	no	Defaults to the public image alias when omitted.

GET/api/mgpt/logs

List logs

List recent generation logs, or fetch a single log when identifier is provided. Required scope: mgpt.logs:read.

Query parameter	Type	Required	Notes
`ownerUserId`	string	yes	Owning user; required for scoping.
`kind`	`text` \| `image`	no	Restrict by log kind.
`identifier`	string	no	Fetch a single log by id.
`requestId`	string	no	Filter by caller correlation id.
`page`	integer >= 1	no	1-indexed page number.
`pageSize`	integer 1-100	no	Items per page.

curl "https://mgpt.continualmi.com/api/mgpt/logs?ownerUserId=$USER_ID&kind=text&pageSize=20" \
  -H "Authorization: Bearer $CONTINUAL_API_KEY"

Errors

Errors return JSON with a detail field and an HTTP status.

{ "detail": "Unsupported model: example-model" }

Status	Meaning	Notes
`400`	Validation error	Missing field, invalid messages, or unsupported model.
`401`	Unauthorized	Missing or invalid bearer token.
`403`	Forbidden	Token lacks the required scope.
`5xx`	Provider / storage failure	Retry according to your client policy.

Scopes

Scope	Endpoints
`mgpt.responses:create`	`POST /chat/completions`
`mgpt.images:create`	`POST /images`
`mgpt.logs:read`	`GET /logs`