Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/cloudwaddie/lmarenabridge/llms.txt

Use this file to discover all available pages before exploring further.

The /api/v1/chat/completions endpoint is the primary way to send messages to LMArena models through the bridge. It follows the OpenAI Chat Completions format, so any client that works with OpenAI’s API can be pointed at LMArena Bridge with minimal changes.

​
Authentication

​
Authorization
string
required
Bearer token. Pass the API key you created in the dashboard: Authorization: Bearer <api_key>. If no API keys are configured on the bridge, pass an empty string or omit the header.

​
Request body

​
model
string
required
The public model ID to use. Retrieve the current list of available IDs from GET /api/v1/models.
​
messages
object[]
required
Array of message objects forming the conversation. Each object must have a role ("system", "user", or "assistant") and a content field.
​
stream
boolean
default:"false"
When true, the response is returned as a Server-Sent Events stream. Each event carries a JSON delta; the stream ends with data: [DONE].
​
temperature
number
Sampling temperature. Passed through to LMArena when provided.
​
max_tokens
number
Maximum number of tokens to generate. Passed through to LMArena when provided.

​
Multi-turn conversations

The bridge maintains a chat session for each API key. When you send a sequence of messages that begins with the same first user message, the bridge routes follow-up turns to the same LMArena session automatically — you do not need to track a session ID yourself. The conversation key is derived from your API key, the model name, and the first user message.
If you change the model or the first user message, a new LMArena session is started. The bridge keeps all sessions in memory; restarting the server clears them.

​
Examples

curl -X POST http://localhost:8000/api/v1/chat/completions \
  -H "Authorization: Bearer your_api_key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o",
    "messages": [
      {"role": "user", "content": "Explain the difference between TCP and UDP."}
    ]
  }'

​
Responses

​
Non-streaming response

200
{
  "id": "chatcmpl-a1b2c3d4e5f6",
  "object": "chat.completion",
  "created": 1716000000,
  "model": "gpt-4o",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "TCP provides reliable, ordered delivery with error checking, while UDP is connectionless and prioritises low latency over reliability."
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 18,
    "completion_tokens": 32,
    "total_tokens": 50
  }
}

​
Streaming response

Each SSE event has the form data: <json>\n\n. The final event is data: [DONE].
SSE stream
: keep-alive

data: {"id":"chatcmpl-a1b2c3d4e5f6","object":"chat.completion.chunk","created":1716000000,"model":"gpt-4o","choices":[{"index":0,"delta":{"role":"assistant","content":""},"finish_reason":null}]}

data: {"id":"chatcmpl-a1b2c3d4e5f6","object":"chat.completion.chunk","created":1716000000,"model":"gpt-4o","choices":[{"index":0,"delta":{"content":"TCP"},"finish_reason":null}]}

data: {"id":"chatcmpl-a1b2c3d4e5f6","object":"chat.completion.chunk","created":1716000000,"model":"gpt-4o","choices":[{"index":0,"delta":{"content":" provides"},"finish_reason":null}]}

data: [DONE]

​
Response fields

​
id
string
Unique identifier for this completion.
​
object
string
Always "chat.completion" for non-streaming responses.
​
model
string
The model that generated the response.
​
choices
object[]
​
usage
object
Token usage counts: prompt_tokens, completion_tokens, total_tokens.

​
Error responses

StatusMeaning
400Invalid JSON, missing model or messages, empty messages array, or prompt exceeds the ~113,000 character limit.
401The Authorization header is missing or the API key is invalid.
403Attempted to use a stealth model (one without a public organisation).
404The requested model was not found. Check GET /api/v1/models for valid IDs.
429Rate limit exceeded for your API key.
503LMArena is unavailable, the model list could not be fetched, or the bridge failed to acquire a reCAPTCHA token.
401
{
  "detail": "Invalid API key"
}
400
{
  "detail": "Prompt too long (120000 characters). LMArena has a character limit of approximately 113567 characters. Please reduce the message size."
}

Build docs developers (and LLMs) love

Get started for freeTalk to us