Chat Completions API

The /v1/chat/completions endpoint allows you to interact with advanced large language models (LLMs) through a unified API, supporting both OpenAI-compatible and Anthropic-compatible models. This endpoint supports conversational AI, function calling, streaming, and multimodal (text+image) input, depending on model capabilities.

Endpoint

POST https://api.llm.vin/v1/chat/completions

Authentication

• API Key (optional):
  You may provide an API key via the Authorization: Bearer ... header.
  • Authenticated users may have access to additional models or higher rate limits.
  • Unauthenticated requests are allowed, but may have restricted model access.

Request Format

{
  "model": "grok-3-mini",
  "messages": [
    {
      "role": "user",
      "content": "Write a one-sentence bedtime story about a unicorn."
    }
  ],
  "temperature": 1.0,
  "max_tokens": 256,
  "stream": false,
  "tools": [],
  "tool_choice": null,
  "stop": null
}

Required Parameters

Optional Parameters

Multimodal Input

• If the model supports image_input, you may include images in the messages array as follows:

  {
    "role": "user",
    "content": [
      { "type": "text", "text": "What is in this image?" },
      { "type": "image_url", "image_url": { "url": "data:image/png;base64,..." } }
    ]
  }
  

Response Format

{
  "id": "chatcmpl-1716151540",
  "object": "chat.completion",
  "created": 1716151540,
  "model": "grok-3-mini",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "Once upon a time, a unicorn danced across the stars and wished you sweet dreams."
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 18,
    "completion_tokens": 22,
    "total_tokens": 40
  }
}

• id: Unique identifier for the chat completion request.
• object: Type of object returned.
• created: Unix timestamp of creation.
• model: Model ID used for the completion.
• choices: Array of response choices, each with:
  • index: Index of the choice.
  • message: The assistant’s reply (role and content).
  • finish_reason: Why the completion stopped (e.g., stop, length, tool_calls).
• usage: Token usage statistics.

Tool Calls (Function Calling)

If the model supports function calling and a tool is invoked, the response may include a tool_calls array in the message:

"tool_calls": [
  {
    "id": "call-abc123",
    "type": "function",
    "function": {
      "name": "get_weather",
      "arguments": "{\"city\": \"Paris\"}"
    }
  }
]

Streaming

If you set "stream": true, the response will be sent as a series of Server-Sent Events (SSE), with each chunk containing a partial completion:

• Each chunk is a JSON object prefixed by data: and followed by two newlines.
• The stream ends with data: [DONE].

Example stream chunk:

data: {"id":"chatcmpl-...","object":"chat.completion.chunk","created":...,"model":"gpt-4.1","choices":[{"index":0,"delta":{"content":"Once upon a time,"},"finish_reason":null}]}

data: {"id":"chatcmpl-...","object":"chat.completion.chunk","created":...,"model":"grok-3-mini","choices":[{"index":0,"delta":{"content":" a unicorn"},"finish_reason":null}]}

data: [DONE]

Available Models

Use the /v1/models endpoint to list available models and their capabilities.

• Only models with chat_completions capability can be used here.
• Some models may support image_input and/or function_calling.

Example Requests

Basic Chat Completion

curl "https://api.llm.vin/v1/chat/completions" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "grok-3-mini",
    "messages": [
      {
        "role": "user",
        "content": "Write a one-sentence bedtime story about a unicorn."
      }
    ]
  }'

Streaming Chat Completion

curl "https://api.llm.vin/v1/chat/completions" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "grok-3-mini",
    "messages": [
      {
        "role": "user",
        "content": "Tell me a joke."
      }
    ],
    "stream": true
  }'

Chat with Function Calling

curl "https://api.llm.vin/v1/chat/completions" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "grok-3-mini",
    "messages": [
      {
        "role": "user",
        "content": "What is the weather in Paris?"
      }
    ],
    "tools": [
      {
        "type": "function",
        "function": {
          "name": "get_weather",
          "description": "Get the current weather for a city.",
          "parameters": {
            "type": "object",
            "properties": {
              "city": { "type": "string" }
            },
            "required": ["city"]
          }
        }
      }
    ]
  }'

Multimodal (Text + Image) Chat

curl "https://api.llm.vin/v1/chat/completions" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "grok-3-mini",
    "messages": [
      {
        "role": "user",
        "content": [
          { "type": "text", "text": "What is in this image?" },
          { "type": "image_url", "image_url": { "url": "data:image/png;base64,..." } }
        ]
      }
    ]
  }'

Error Handling

The API returns standard HTTP status codes:

Error responses include a JSON object:

{
  "error": {
    "message": "Model 'grok-3-mini' not found",
    "type": "invalid_request_error",
    "code": "model_not_found"
  }
}

Rate Limits

• Chat completions: 500 requests per day and 10 requests per minute per IP.
• Other endpoints: 50,000 requests per day per IP.
• Limits may be higher for authenticated users.

Notes

• Use /v1/models to discover available models and their capabilities.
• Function calling and tool use are only supported by models with those capabilities enabled.
• Streaming responses are available by setting stream: true.
• Multimodal (image) input is only supported by models with image_input capability.
• If you provide an API key, you may have access to more models or higher rate limits.
• All requests and errors are logged for security and debugging.