Chat (OpenAI-compatible)

POST https://gateway.mindproxy.ai/v1/chat/completions

Compatible with OpenAI Chat Completions; supports streaming, tool calling and multimodal (depending on the model).

Request

curl https://gateway.mindproxy.ai/v1/chat/completions \
  -H "Authorization: Bearer $TT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o-mini",
    "messages": [
      {"role": "system", "content": "You are a concise assistant."},
      {"role": "user", "content": "Introduce Hangzhou in one sentence."}
    ],
    "temperature": 0.7,
    "max_tokens": 256
  }'

Common fields:

Field	Description
model	Public model ID, see Models & plans
messages	Conversation messages (role: system / user / assistant / tool)
stream	true enables streaming (SSE)
temperature / top_p	Sampling parameters
max_tokens	Max output tokens
tools / tool_choice	Tool calling, see Streaming & tools

Other OpenAI fields are generally accepted as-is.

Response (non-streaming)

{
  "id": "chatcmpl-...",
  "object": "chat.completion",
  "model": "gpt-4o-mini",
  "choices": [
    {
      "index": 0,
      "message": {"role": "assistant", "content": "Hangzhou is the capital of Zhejiang, famous for West Lake."},
      "finish_reason": "stop"
    }
  ],
  "usage": {"prompt_tokens": 24, "completion_tokens": 14, "total_tokens": 38}
}

usage is the billing basis (per token). Response headers include X-TT-Request-ID etc., see Auth & Base URL.

Streaming

Add "stream": true to get SSE: data: {...} chunks (deltas), ending with data: [DONE]. SDKs wrap this — see Python / Node.js.

Other OpenAI-compatible capabilities

The same prefix also offers: /v1/embeddings, /v1/responses, /v1/images/generations, /v1/images/edits, /v1/audio/transcriptions, /v1/audio/speech. Images: Image generation.