◐llm.market

API REFERENCE

Chat completions

Generate chat responses through llm.market's OpenAI-compatible gateway and active model catalog.

llm.market exposes the supported OpenAI-compatible endpoints shown here. Use the live model catalog before choosing a model ID.

Authentication

Example
Authorization: Bearer sk-YOUR_LLM_MARKET_KEY

POST /v1/chat/completions

OpenAI-compatible Chat Completions. Works with the official openai SDK by overriding base_url to https://llm.market/v1.

Request body

ParameterTypeRequiredDescription
modelstringRequiredCatalog id from GET /v1/models.
messagesarrayRequiredConversation history.
max_tokens / max_completion_tokensintegerOptionalOutput cap. gpt-5* prefers max_completion_tokens; both are forwarded.
reasoning_effortstringOptionalOpenAI-style none|low|medium|high|xhigh. Forwarded as-is on GPT channels.
thinkingobjectOptionalClaude Messages only: {type, budget_tokens} or adaptive effort. Use POST /v1/messages.
temperaturefloatOptional0–2 when the upstream model allows it. Ignored on some gpt-5/o-series models.
streambooleanOptionalSSE. Chat Completions requests include stream_options.include_usage.
toolsarrayOptionalOpenAI function tools on /v1/chat/completions; Claude tool_use on /v1/messages.

Basic example

Example
curl https://llm.market/v1/chat/completions \
  -H "Authorization: Bearer sk-YOUR_LLM_MARKET_KEY" \
  -H "Content-Type: application/json" \
   -d '{
     "model": "MODEL_FROM_V1_MODELS",
    "messages": [
      {"role": "system", "content": "You are a helpful assistant."},
      {"role": "user", "content": "What is the capital of France?"}
    ],
    "max_tokens": 200,
    "temperature": 0.7
  }'

Response shape

Example
{
  "id": "chatcmpl-abc123",
  "object": "chat.completion",
  "created": 1710000000,
  "model": "MODEL_FROM_V1_MODELS",
  "choices": [{"index":0,"message":{"role":"assistant","content":"The capital of France is Paris."},"finish_reason":"stop"}],
  "usage": {"prompt_tokens":20,"completion_tokens":8,"total_tokens":28}
}

Reasoning & thinking

Do not send Claude thinking blocks on /v1/chat/completions. GPT models take reasoning_effort on Chat Completions. Claude Code should call POST /v1/messages with the same API key (Authorization Bearer or x-api-key) and ANTHROPIC_BASE_URL=https://llm.market — not https://llm.market/v1.

Example
{
  "env": {
    "ANTHROPIC_API_KEY": "sk-YOUR_LLM_MARKET_KEY",
    "ANTHROPIC_BASE_URL": "https://llm.market",
    "ANTHROPIC_MODEL": "claude-sonnet-5"
  }
}

Vision & image input

Example
{"role":"user","content":[{"type":"text","text":"What is in this image?"},{"type":"image_url","image_url":{"url":"https://example.com/cat.jpg"}}]}

Need a key?

Create one in the dashboard. Discord for llm.market is coming soon.

Get a keyDiscord