API Overview

Kouri Ai provides API services fully compatible with OpenAI API format, while also supporting Anthropic, Gemini and other protocols, allowing you to easily call dozens of flagship models including GPT, Claude, Gemini, DeepSeek, and more.

Endpoints

OpenAI Compatible Endpoints

Endpoint Type	URL	Description
Chat Completions	`https://api.kourichat.com/v1/chat/completions`	Chat completion API, for most models
Responses	`https://api.kourichat.com/v1/responses`	Response API, required for reasoning models
Standard Endpoint	`https://api.kourichat.com/v1`	Recommended for SDKs
Base Endpoint	`https://api.kourichat.com`	For some applications

Important: Some models like gpt-5.2-pro and o3-pro only support the Response API, not Chat Completions. Please choose the correct API based on model requirements.

Other Protocol Endpoints

Protocol	Endpoint URL	Description
Anthropic Protocol	`https://api.kourichat.com/v1/messages`	Claude native protocol
Gemini Protocol	`https://api.kourichat.com/v1beta`	Gemini native protocol

Model Compatibility: Both Anthropic and Gemini protocol endpoints support calling all models (not limited to Claude or Gemini), while the OpenAI protocol's Responses endpoint only supports specific reasoning models.

Authentication

All API requests require authentication via API token. You can create and manage tokens in the Console.

HTTP Header Authentication

Add the Authorization header to your request:

Authorization: Bearer sk-xxxxxxxxxxxxxxxx

Complete Request Example

curl https://api.kourichat.com/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer sk-xxxxxxxx" \
  -d '{
    "model": "gpt-4o",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

Security Notice: Keep your API token safe. Never expose it in client-side code, public repositories, or logs.

Request Format

All API requests use JSON format:

Content-Type: application/json
Method: POST (chat endpoints), GET (query endpoints)

Basic Request Structure

{
  "model": "model-name",
  "messages": [
    {"role": "system", "content": "System prompt"},
    {"role": "user", "content": "User message"}
  ],
  "temperature": 0.7,
  "max_tokens": 2048,
  "stream": false
}

Common Parameters

Parameter	Type	Required	Description
`model`	string	✅	Model name, e.g., `gpt-4o`, `claude-sonnet-4-20250514`
`messages`	array	✅	Message list with roles and content
`temperature`	number	❌	Randomness, 0-2, default 1
`max_tokens`	integer	❌	Maximum output tokens
`stream`	boolean	❌	Enable streaming, default false
`top_p`	number	❌	Nucleus sampling, 0-1

Response Format

Standard Response

{
  "id": "chatcmpl-xxxxxxxx",
  "object": "chat.completion",
  "created": 1234567890,
  "model": "gpt-4o",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "Hello! How can I help you?"
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 10,
    "completion_tokens": 15,
    "total_tokens": 25
  }
}

Streaming Response

With stream: true, responses are returned as Server-Sent Events (SSE):

data: {"id":"chatcmpl-xxx","choices":[{"delta":{"content":"Hello"},"index":0}]}

data: {"id":"chatcmpl-xxx","choices":[{"delta":{"content":"!"},"index":0}]}

data: [DONE]

Error Codes

HTTP Status	Error Type	Description
400	Bad Request	Invalid request parameters
401	Unauthorized	Invalid or missing token
403	Forbidden	Access denied
404	Not Found	Endpoint or model not found
429	Too Many Requests	Rate limit exceeded
500	Internal Server Error	Server error
503	Service Unavailable	Service temporarily unavailable

Error Response Example

{
  "error": {
    "message": "Invalid API key provided",
    "type": "invalid_request_error",
    "code": "invalid_api_key"
  }
}

Universal Model Access

Kouri Ai provides universal model access, allowing you to call all chat models through the standard ChatCompletion endpoint, including:

OpenAI Series: GPT-4o, GPT-4, o1, o3, etc.
Anthropic Series: Claude Sonnet, Claude Opus, etc.
Google Series: Gemini 2.5 Flash, Gemini 2.5 Pro, etc.
Other Models: DeepSeek, Qwen, etc.

Kouri Ai handles protocol conversion automatically. You can use /v1/chat/completions to call all models without worrying about underlying protocol differences.

Rate Limits

Request Rate: Varies by token type and account level
Request Timeout: 5 minutes for normal requests, longer for complex reasoning
Max Message Length: Depends on the model's context window

For higher quotas, please contact support for enterprise plans.

API Overview

On this page