API Manual

TK-NovaLink provides a complete OpenAI-compatible API. This document lists all available endpoints, parameters, and response formats.

API Endpoints

API	Path	Method
Chat Completions	/v1/chat/completions	POST
Text Completions	/v1/completions	POST
Embeddings	/v1/embeddings	POST
Image Generation	/v1/images/generations	POST
Audio Transcription	/v1/audio/transcriptions	POST
List Models	/v1/models	GET

Authentication

All API requests require a Bearer Token in the HTTP Header:

http

Authorization: Bearer sk-xxxxxxxxxxxxxxxxxxxxxxxx

Chat Completion Parameters

The chat completion endpoint supports the following parameters:

Parameter	Type	Required	Description
model	string	Yes	Model name, e.g. gpt-4o
messages	array	Yes	Array of chat messages
temperature	number	No	Sampling temperature, 0-2, default 1
max_tokens	integer	No	Max tokens to generate
stream	boolean	No	Whether to stream the response, default false
top_p	number	No	Nucleus sampling parameter, 0-1

Request Example

json

{
  "model": "gpt-4o",
  "messages": [
    {"role": "system", "content": "You are a helpful assistant."},
    {"role": "user", "content": "Hello!"}
  ],
  "max_tokens": 1024,
  "temperature": 0.7,
  "stream": false
}

Response Example

json

{
  "id": "chatcmpl-xxx",
  "object": "chat.completion",
  "created": 1700000000,
  "model": "gpt-4o",
  "choices": [{
    "index": 0,
    "message": {
      "role": "assistant",
      "content": "Hello! How can I help you today?"
    },
    "finish_reason": "stop"
  }],
  "usage": {
    "prompt_tokens": 15,
    "completion_tokens": 8,
    "total_tokens": 23
  }
}

Error Codes

When a request fails, the API returns an error code and message:

Error Code	HTTP Status	Description
invalid_api_key	401	Invalid API key
insufficient_quota	429	Insufficient quota
model_not_found	404	Model not found
invalid_request_error	400	Invalid request parameters
server_error	500	Internal server error
rate_limit_exceeded	429	Rate limit exceeded

Rate Limits

TK-NovaLink enforces rate limits to ensure service quality.

Rate limits (RPM/TPM) vary by model, token group, route, and account settings. Exceeding the limit returns HTTP 429; we recommend retrying with exponential backoff.

Pricing & Billing