### Task Started Event Example

Source: https://platform.minimax.io/docs/api-reference/speech-t2a-websocket

Example payload indicating that the Speech-to-Text task has commenced. This event includes session and trace identifiers for monitoring.

```json
{
  "session_id": "xxxx",
  "event": "task_started",
  "trace_id": "0303a2882bf18235ae7a809ae0f3cca7",
  "base_resp": {
    "status_code": 0,
    "status_msg": "success"
  }
}
```

--------------------------------

### Install OpenAI SDK (Node.js)

Source: https://platform.minimax.io/docs/api-reference/text-openai-api

Install the official OpenAI Node.js package using npm.

```bash
npm install openai
```

--------------------------------

### Install OpenAI SDK (Python)

Source: https://platform.minimax.io/docs/api-reference/text-openai-api

Install the official OpenAI Python package using pip.

```bash
pip install openai
```

--------------------------------

### Install AI SDK and MiniMax Provider (npm)

Source: https://platform.minimax.io/docs/api-reference/text-ai-sdk

Install the AI SDK and the Vercel MiniMax AI provider using npm.

```bash
npm install ai vercel-minimax-ai-provider
```

--------------------------------

### Install AI SDK and MiniMax Provider (pnpm)

Source: https://platform.minimax.io/docs/api-reference/text-ai-sdk

Install the AI SDK and the Vercel MiniMax AI provider using pnpm.

```bash
pnpm add ai vercel-minimax-ai-provider
```

--------------------------------

### Install Anthropic SDK for Node.js

Source: https://platform.minimax.io/docs/api-reference/text-anthropic-api

Install the official Anthropic SDK for Node.js using npm.

```bash
npm install @anthropic-ai/sdk
```

--------------------------------

### Bearer Authentication Example

Source: https://platform.minimax.io/docs/api-reference/image-generation-i2i

Example of how to use Bearer authentication with the API key.

```http
HTTP: Bearer Auth

- Security Scheme Type: http

- HTTP Authorization Scheme: `Bearer API_key`, can be found in [Account Management>API Keys](https://platform.minimax.io/user-center/basic-information/interface-key).
```

--------------------------------

### Install Anthropic SDK for Python

Source: https://platform.minimax.io/docs/api-reference/text-anthropic-api

Install the official Anthropic SDK for Python using pip.

```bash
pip install anthropic
```

--------------------------------

### Example Response for List Models

Source: https://platform.minimax.io/docs/api-reference/models/openai/list-models

This example demonstrates the structure of a successful response when listing available models. It includes model IDs, creation timestamps, and ownership information.

```json
{
  "object": "list",
  "data": [
    {
      "id": "MiniMax-M3",
      "object": "model",
      "created": 1780272000,
      "owned_by": "minimax"
    },
    {
      "id": "MiniMax-M2.7",
      "object": "model",
      "created": 1773799200,
      "owned_by": "minimax"
    },
    {
      "id": "MiniMax-M2.5",
      "object": "model",
      "created": 1770948000,
      "owned_by": "minimax"
    }
  ]
}
```

--------------------------------

### Lyrics Generation Request Example

Source: https://platform.minimax.io/docs/api-reference/lyrics-generation

Example of a request payload for generating a full song. Ensure the 'mode' is set to 'write_full_song' and provide a descriptive 'prompt'.

```yaml
mode: write_full_song
prompt: A cheerful love song about a summer day at the beach
```

--------------------------------

### Cover Preprocess Request Example

Source: https://platform.minimax.io/docs/api-reference/music-cover-preprocess

Example of a request body for the Music Cover Preprocess API, specifying the model as 'music-cover' and providing a URL for the reference audio.

```json
{
  "model": "music-cover",
  "audio_url": "https://example.com/song.mp3"
}
```

--------------------------------

### Estimate Input Tokens Response Example

Source: https://platform.minimax.io/docs/api-reference/responses-input-tokens

This example shows a successful response from the Estimate Input Tokens API, indicating the calculated input token count.

```json
{
  "object": "response.input_tokens",
  "input_tokens": 588
}
```

--------------------------------

### Estimate Input Tokens Request Example

Source: https://platform.minimax.io/docs/api-reference/responses-input-tokens

This example shows the structure of a request to estimate input tokens. It includes the model name, conversation history with messages and tool definitions, and a system instruction.

```json
{
  "model": "MiniMax-M3",
  "input": [
    {
      "type": "message",
      "role": "user",
      "content": "> Please implement a generic quicksort algorithm in Python\nwith these requirements: 1) in-place sorting to save\nmemory; 2) three-way partitioning to handle duplicate\nelements; 3) switch to insertion sort for small\nsubarrays; 4) include complete unit tests. Finally,\nexplain the advantage of three-way partitioning over the\nclassic Lomuto scheme when keys repeat."
    }
  ],
  "tools": [
    {
      "type": "function",
      "name": "search_docs",
      "description": "> Search the official documentation of the Python standard\nlibrary or a third-party package",
      "parameters": {
        "type": "object",
        "properties": {
          "library": {
            "type": "string",
            "description": "Library name, e.g. `typing`, `itertools`"
          },
          "query": {
            "type": "string",
            "description": "Search keywords"
          }
        },
        "required": [
          "library",
          "query"
        ]
      }
    },
    {
      "type": "function",
      "name": "run_python",
      "description": "> Execute Python code in a sandbox and return stdout /\nerror",
      "parameters": {
        "type": "object",
        "properties": {
          "code": {
            "type": "string",
            "description": "Python code to execute"
          },
          "timeout_seconds": {
            "type": "integer",
            "description": "Execution timeout in seconds",
            "default": 10
          }
        },
        "required": [
          "code"
        ]
      }
    }
  ]
}
```

--------------------------------

### Python: Start Speech Synthesis Task

Source: https://platform.minimax.io/docs/api-reference/speech-t2a-websocket

Sends the 'task_start' event to initiate a speech synthesis task. Requires a successful WebSocket connection and returns True if the task starts successfully, False otherwise.

```python
async def start_task(websocket):
    """Start task"""
    try:
        await websocket.send(json.dumps({"event": "task_start", "model": "speech-2.8-hd", "voice_setting": {"voice_id": "Chinese (Mandarin)_Lyrical_Voice"}}))
        response = await websocket.recv()
        if json.loads(response).get("event") == "task_started":
            return True
        return False
    except Exception as e:
        print(f"Error: {e}")
        return False
```

--------------------------------

### OpenAPI Specification for Query Video Generation Task

Source: https://platform.minimax.io/docs/api-reference/video-generation-query

This OpenAPI specification defines the GET /v1/query/video_generation endpoint for querying the status of a video generation task. It includes request parameters, response schemas, and example responses.

```yaml
openapi: 3.1.0
info:
  title: MiniMax API
  description: MiniMax video generation and file management API
  license:
    name: MIT
  version: 1.0.0
servers:
  - url: https://api.minimax.io
security:
  - bearerAuth: []
paths:
  /v1/query/video_generation:
    get:
      tags:
        - Video
      summary: Query Video Generation Task
      operationId: queryVideoGenerationTask
      parameters:
        - name: task_id
          in: query
          required: true
          description: >-
            The task ID to query. Only tasks created under the current account
            can be queried.
          schema:
            type: string
      responses:
        '200':
          description: ''
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/QueryVideoGenerationTaskResp'
components:
  schemas:
    QueryVideoGenerationTaskResp:
      type: object
      properties:
        task_id:
          type: string
          description: The queried task ID.
        status:
          $ref: '#/components/schemas/VideoProcessStatus'
        file_id:
          type: string
          description: >-
            Returned when the task succeeds. Represents the file ID of the
            generated video.
        video_width:
          type: integer
          description: >-
            Returned when the task succeeds. The width of the generated video
            (in pixels).
        video_height:
          type: integer
          description: >-
            Returned when the task succeeds. The height of the generated video
            (in pixels).
        base_resp:
          $ref: '#/components/schemas/QueryVideoGenerationTaskBaseResp'
      example:
        task_id: '176843862716480'
        status: Success
        file_id: '176844028768320'
        video_width: 1920
        video_height: 1080
        base_resp:
          status_code: 0
          status_msg: success
    VideoProcessStatus:
      type: string
      enum:
        - Preparing
        - Queueing
        - Processing
        - Success
        - Fail
      description: |- 
        The current status of the task. Possible values:
        - `Preparing` – Preparing
        - `Queueing` – In queue
        - `Processing` – Generating
        - `Success` – Completed successfully
        - `Fail` – Failed
    QueryVideoGenerationTaskBaseResp:
      type: object
      properties:
        status_code:
          type: integer
          description: |- 
            The status codes are as follows:
            - `0`, Request successful;
            - `1002`,  Rate limit triggered, retry later
            - `1004`, Authentication failed, check API Key
            - `1026`, Contains sensitive content in the input
            - `1027`, Contains sensitive content in the generated video

             For more information, please refer to the [Error Code Reference](/api-reference/errorcode).
        status_msg:
          type: string
          description: Status message. Returns `success` when successful.
  securitySchemes:
    bearerAuth:
      type: http
      scheme: bearer
      bearerFormat: JWT
      description: >-
        `HTTP: Bearer Auth`

        - Security Scheme Type: http

        - HTTP Authorization Scheme: `Bearer API_key`, can be found in [Account
        Management>API
        Keys](https://platform.minimax.io/user-center/basic-information/interface-key).
```

--------------------------------

### Callback Handling

Source: https://platform.minimax.io/docs/api-reference/video-generation-t2v

This section provides an example of how to set up a server to handle callback requests from the MiniMax platform for video generation tasks. It includes validation and status update handling.

```APIDOC
## Callback URL Handling

### Description
MiniMax sends `POST` requests to the configured `callback_url` to provide asynchronous task status updates. The server must first validate the callback by echoing a `challenge` field, and then process status updates.

### Callback `status` values:
- `"processing"` – Task in progress
- `"success"` – Task completed successfully
- `"failed"` – Task failed

### Example Server Implementation (Python/FastAPI)
```python
from fastapi import FastAPI, HTTPException, Request
import json

app = FastAPI()

@app.post("/get_callback")
async def get_callback(request: Request):
    try:
        json_data = await request.json()
        challenge = json_data.get("challenge")
        if challenge is not None:
            # Validation request, echo back challenge
            return {"challenge": challenge}
        else:
            # Status update request, handle accordingly
            # Example status update payload:
            # {
            #     "task_id": "115334141465231360",
            #     "status": "success",
            #     "file_id": "205258526306433",
            #     "base_resp": {
            #         "status_code": 0,
            #         "status_msg": "success"
            #     }
            # }
            print(f"Received status update: {json_data}") # Process the status update
            return {"status": "success"}
    except Exception as e:
        raise HTTPException(status_code=500, detail=str(e))

if __name__ == "__main__":
    import uvicorn
    uvicorn.run(
        app,  # Required
        host="0.0.0.0",  # Required
        port=8000,  # Required, port can be customized
    )
```
```

--------------------------------

### Simple Text Input Example

Source: https://platform.minimax.io/docs/api-reference/responses-create

Demonstrates a simple text input for generating a response.

```json
{
  "model": "MiniMax-M3",
  "input": "Hello!"
}
```

--------------------------------

### Image Generation Request Example

Source: https://platform.minimax.io/docs/api-reference/image-generation-i2i

Example of a request to generate images with specified parameters including subject reference.

```json
{
  "model": "image-01",
  "prompt": "A girl looking into the distance from a library window",
  "aspect_ratio": "16:9",
  "subject_reference": [
    {
      "type": "character",
      "image_file": "https://cdn.hailuoai.com/prod/2025-08-12-17/video_cover/1754990600020238321-411603868533342214-cover.jpg"
    }
  ],
  "n": 2
}
```

--------------------------------

### TimbreWeights Example

Source: https://platform.minimax.io/docs/api-reference/speech-t2a-http

Example of how to configure timbre weights for voice synthesis.

```json
            "timbre_weights": [
              {
                "voice_id": "female-chengshu",
                "weight": 30
              },
              {
                "voice_id": "female-tianmei",
                "weight": 70
              }
            ]
```

--------------------------------

### Function Call Example

Source: https://platform.minimax.io/docs/api-reference/responses-create

Demonstrates how to define tools for function calls, allowing the model to interact with external functions.

```json
{
  "model": "MiniMax-M3",
  "input": "What is the weather in Boston today?",
  "tools": [
    {
      "type": "function",
      "name": "get_current_weather",
      "description": "Get the current weather in a given location",
      "parameters": {
        "type": "object",
        "properties": {
          "location": {
            "type": "string",
            "description": "The city and state, e.g. San Francisco, CA"
          },
          "unit": {
            "type": "string",
            "enum": [
              "celsius",
              "fahrenheit"
            ]
          }
        },
        "required": [
          "location",
          "unit"
        ]
      }
    }
  ]
}
```

--------------------------------

### Connected Success Event Example

Source: https://platform.minimax.io/docs/api-reference/speech-t2a-websocket

Example payload for a successful connection event. Use this to confirm the WebSocket connection is established.

```json
{
  "session_id": "xxxx",
  "event": "connected_success",
  "trace_id": "0303a2882bf18235ae7a809ae0f3cca7",
  "base_resp": {
    "status_code": 0,
    "status_msg": "success"
  }
}
```

--------------------------------

### Streaming Output Example

Source: https://platform.minimax.io/docs/api-reference/responses-create

Enables streaming output for real-time response generation.

```json
{
  "model": "MiniMax-M3",
  "input": "Hello!",
  "stream": true
}
```

--------------------------------

### Image Generation Response Example

Source: https://platform.minimax.io/docs/api-reference/image-generation-i2i

Example of a successful response from the image generation API, including image URLs and metadata.

```json
{
  "id": "03ff3cd0820949eb8a410056b5f21d38",
  "data": {
    "image_urls": [
      "XXX",
      "XXX",
      "XXX"
    ]
  },
  "metadata": {
    "failed_count": "0",
    "success_count": "3"
  },
  "base_resp": {
    "status_code": 0,
    "status_msg": "success"
  }
}
```

--------------------------------

### Task Started Event

Source: https://platform.minimax.io/docs/api-reference/speech-t2a-websocket

Notification that the Text-to-Speech (T2A) task has started. This event provides initial session and trace information.

```APIDOC
## Task Started Event

### Description
Notification that T2A task has started. This event provides initial session and trace information.

### Event Type
task_started

### Payload
```json
{
  "session_id": "xxxx",
  "event": "task_started",
  "trace_id": "0303a2882bf18235ae7a809ae0f3cca7",
  "base_resp": {
    "status_code": 0,
    "status_msg": "success"
  }
}
```

### Response Example
```json
{
  "session_id": "xxxx",
  "event": "task_started",
  "trace_id": "0303a2882bf18235ae7a809ae0f3cca7",
  "base_resp": {
    "status_code": 0,
    "status_msg": "success"
  }
}
```
```

--------------------------------

### Timbre Weights Configuration Example

Source: https://platform.minimax.io/docs/api-reference/speech-t2a-websocket

Configure voice mixing for synthesis by specifying voice IDs and their corresponding weights. Up to 4 voices can be mixed, with higher weights increasing similarity to that voice.

```json
"timbre_weights": [
  {
    "voice_id": "female-chengshu",
    "weight": 30
  },
  {
    "voice_id": "female-tianmei",
    "weight": 70
  }
]
```

--------------------------------

### Call MiniMax API using Anthropic SDK in Python

Source: https://platform.minimax.io/docs/api-reference/text-anthropic-api

Example of initializing the Anthropic client and making a message creation request to a MiniMax model. Ensure the complete model response is appended to the conversation history for multi-turn conversations.

```python
import anthropic

client = anthropic.Anthropic()

message = client.messages.create(
    model="MiniMax-M3",
    max_tokens=1000,
    system="You are a helpful assistant.",
    messages=[
        {
            "role": "user",
            "content": [
                {
                    "type": "text",
                    "text": "Hi, how are you?"
                }
            ]
        }
    ]
)

for block in message.content:
    if block.type == "thinking":
        print(f"Thinking:\n{block.thinking}\n")
    elif block.type == "text":
        print(f"Text:\n{block.text}\n")
```

--------------------------------

### OpenAPI Specification for Get Voice

Source: https://platform.minimax.io/docs/api-reference/voice-management-get

This OpenAPI specification defines the structure and parameters for the POST /v1/get_voice API endpoint, used to retrieve voice information.

```yaml
openapi: 3.1.0
info:
  title: MiniMax Voice Management API
  description: MiniMax Voice Management API with support for getting and deleting voices
  license:
    name: MIT
  version: 1.0.0
servers:
  - url: https://api.minimax.io
security:
  - bearerAuth: []
paths:
  /v1/get_voice:
    post:
      tags:
        - Voice
      summary: Get Voice
      operationId: getVoice
      parameters:
        - name: Content-Type
          in: header
          required: true
          description: >-
            The media type of the request body. Must be set to
            `application/json` to ensure the data is sent in JSON format.
          schema:
            type: string
            enum:
              - application/json
            default: application/json
      requestBody:
        description: ''
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/GetVoiceReq'
        required: true
      responses:
        '200':
          description: ''
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/GetVoiceResp'
components:
  schemas:
    GetVoiceReq:
      type: object
      required:
        - voice_type
      properties:
        voice_type:
          type: string
          description: >-
            The type of voice to query. Supported values:

            - `system`: System voices

            - `voice_cloning`: Quick cloned voices (available only after being
            successfully used for speech synthesis)

            - `voice_generation`: Voices generated via the text-to-voice API
            (available only after being successfully used for speech synthesis)

            - `all`: All of the above
          enum:
            - system
            - voice_cloning
            - voice_generation
            - all
      example:
        voice_type: all
    GetVoiceResp:
      type: object
      properties:
        system_voice:
          type: array
          items:
            $ref: '#/components/schemas/SystemVoiceInfo'
          description: Contains system-defined voices
        voice_cloning:
          type: array
          items:
            $ref: '#/components/schemas/VoiceCloningInfo'
          description: Contains data of quick cloned voices
        voice_generation:
          type: array
          items:
            $ref: '#/components/schemas/VoiceGenerationInfo'
          description: Contains data of voices generated via the text-to-voice API
        base_resp:
          $ref: '#/components/schemas/BaseResp'
      example:
        system_voice:
          - voice_id: Chinese (Mandarin)_Reliable_Executive
            description:
              - >-
                A steady and reliable male executive voice in standard Mandarin,
                conveying a trustworthy impression.
            voice_name: Steady Executive
            created_time: '1970-01-01'
          - voice_id: Chinese (Mandarin)_News_Anchor
            description:
              - >-
                A professional middle-aged female news anchor voice in standard
                Mandarin, with a broadcast-style tone.
            voice_name: News Anchor (Female)
            created_time: '1970-01-01'
        voice_cloning:
          - voice_id: test12345
            description: []
            created_time: '2025-08-20'
          - voice_id: test12346
            description: []
            created_time: '2025-08-21'
        voice_generation:
          - voice_id: ttv-voice-2025082011321125-2uEN0X1S
            description: []
            created_time: '2025-08-20'
          - voice_id: ttv-voice-2025082014225025-ZCQt0U0k
            description: []
            created_time: '2025-08-20'
        base_resp:
          status_code: 0
          status_msg: success
    SystemVoiceInfo:
      type: object
      properties:
        voice_id:
          type: string
          description: The voice ID.
        voice_name:
          type: string
          description: The voice display name (not the API call ID).
        description:
          type: array
          items:
            type: string
          description: The voice description.
    VoiceCloningInfo:
      type: object
      properties:
        voice_id:
          type: string
          description: Quick cloned voice ID.
        description:
          type: array
          items:
            type: string
          description: The description provided during cloning.
        created_time:
          type: string
          description: Creation date in `yyyy-mm-dd` format.
    VoiceGenerationInfo:
      type: object
      properties:
        voice_id:
          type: string
          description: The generated voice ID.
        description:
          type: array
          items:
            type: string

```