API Documentation

Get started with the KiroGate API

🔑

Authentication

All API requests must include an API key in the header. Three authentication modes are supported:

Mode 1: User API Key (sk-xxx format) ⭐ Recommended

# OpenAI format
Authorization: Bearer sk-xxxxxxxxxxxxxxxx

# Anthropic format
x-api-key: sk-xxxxxxxxxxxxxxxx

Generate one in the user center after login. Uses your added tokens or the public token pool automatically.

Mode 2: Combined Mode (user-supplied REFRESH_TOKEN)

# OpenAI format
Authorization: Bearer YOUR_PROXY_API_KEY:YOUR_REFRESH_TOKEN

# Anthropic format
x-api-key: YOUR_PROXY_API_KEY:YOUR_REFRESH_TOKEN

Mode 3: Simple Mode (uses server-configured REFRESH_TOKEN)

# OpenAI format
Authorization: Bearer YOUR_PROXY_API_KEY

# Anthropic format
x-api-key: YOUR_PROXY_API_KEY

💡 Recommended Usage

  • • Regular users: After login, generate an API key in sk-xxx format - the simplest option.
  • • Self-hosted users: Use combined mode with your own REFRESH_TOKEN - no server config needed.
  • • Cache optimization: Each user's auth info is cached (up to 100 users) for improved performance.
📡

Endpoints

GET
/

Health check endpoint

GET
/health

Detailed health check; returns token status and cache info

GET
/v1/models

List available models (authentication required)

POST
/v1/chat/completions

OpenAI-compatible chat completions API (authentication required)

POST
/v1/messages

Anthropic-compatible messages API (authentication required)

💡

Usage Examples

🐍 OpenAI SDK (Python) 

from openai import OpenAI

client = OpenAI(
    base_url="http://localhost:8000/v1",
    api_key="YOUR_PROXY_API_KEY"
)

response = client.chat.completions.create(
    model="claude-sonnet-4-5",
    messages=[{"role": "user", "content": "Hello!"}],
    stream=True
)

for chunk in response:
    print(chunk.choices[0].delta.content, end="")

🤖 Anthropic SDK (Python) 

import anthropic

client = anthropic.Anthropic(
    base_url="http://localhost:8000",
    api_key="YOUR_PROXY_API_KEY"
)

message = client.messages.create(
    model="claude-sonnet-4-5",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Hello!"}]
)

print(message.content[0].text)

$ cURL 

curl http://localhost:8000/v1/chat/completions \
  -H "Authorization: Bearer YOUR_PROXY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-sonnet-4-5",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'
🛠️

CLI Tools Integration

Auto-generated configuration for popular AI coding tools. The base URL is pre-filled with this server's address.

Claude Code

Claude Code is Anthropic's official CLI tool. It uses the Anthropic API format natively.

Step 1: Set environment variables


                
              

            

            

              
Step 2: Run Claude Code

              

                
claude

              

            

            

              
Note: Claude Code will automatically use the configured base URL and API key. All Claude models available on this gateway will work.

            

          


          
          

            

              
OpenCode

              
OpenCode is an open-source AI coding assistant that supports OpenAI-compatible APIs.

            

            

              
Step 1: Set environment variables

              

                

                
              

            

            

              
Step 2: Configure in opencode.json

              

                

                
              

            

            

              
Step 3: Run OpenCode

              
opencode

            

          


          
          

            

              
OpenClaw

              
OpenClaw is a lightweight CLI tool for interacting with OpenAI-compatible APIs.

            

            

              
Step 1: Set environment variables

              

                

                
              

            

            

              
Step 2: Run OpenClaw

              
openclaw

            

          


          
          

            

              
PicoClaw

              
PicoClaw is a minimal AI coding assistant with OpenAI-compatible API support.

            

            

              
Step 1: Set environment variables

              

                

                
              

            

            

              
Step 2: Run PicoClaw

              
picoclaw

            

          


          
          

            

              
Hermes Agent

              
Hermes Agent is an AI coding agent that supports the Anthropic API format.

            

            

              
Step 1: Set environment variables

              

                

                
              

            

            

              
Step 2: Run Hermes

              
hermes

            

          


          
          

            

              
Qwen Code

              
Qwen Code is an AI coding assistant that supports OpenAI-compatible APIs.

            

            

              
Step 1: Set environment variables

              

                

                
              

            

            

              
Step 2: Run Qwen Code

              
qwen-code

            

          


          
          

            
Available Models

            
Use any of these model IDs in your CLI tool configuration:

            

              autoclaude-opus-4-7claude-opus-4-6claude-opus-4-5claude-sonnet-4-6claude-sonnet-4-5claude-sonnet-4claude-haiku-4-5deepseek-3-2minimax-m2-5minimax-m2-1glm-5qwen3-coder-nextcommandcode/claude-fable-5commandcode/claude-haiku-4-5commandcode/claude-opus-4-6commandcode/claude-opus-4-7commandcode/claude-sonnet-4-6commandcode/deepseek-v4-flashcommandcode/deepseek-v4-procommandcode/gemini-3.1-flash-litecommandcode/gemini-3.5-flashcommandcode/glm-5commandcode/glm-5.1commandcode/gpt-5.3-codexcommandcode/gpt-5.4commandcode/gpt-5.4-minicommandcode/gpt-5.5commandcode/kimi-k2.5commandcode/kimi-k2.6commandcode/kimi-k2.7-codecommandcode/mimo-v2.5commandcode/mimo-v2.5-procommandcode/minimax-m2.5commandcode/minimax-m2.7commandcode/minimax-m3commandcode/nemotron-3-ultracommandcode/qwen-3.6-max-previewcommandcode/qwen-3.6-pluscommandcode/qwen-3.7-maxcommandcode/qwen-3.7-pluscommandcode/step-3.5-flashcommandcode/step-3.7-flashcommandcode/taste-1
            

          

        


        

          

            
🤖

            
Available Models

          

          

            
auto
claude-opus-4-7
claude-opus-4-6
claude-opus-4-5
claude-sonnet-4-6
claude-sonnet-4-5
claude-sonnet-4
claude-haiku-4-5
deepseek-3-2
minimax-m2-5
minimax-m2-1
glm-5
qwen3-coder-next
commandcode/claude-fable-5
commandcode/claude-haiku-4-5
commandcode/claude-opus-4-6
commandcode/claude-opus-4-7
commandcode/claude-sonnet-4-6
commandcode/deepseek-v4-flash
commandcode/deepseek-v4-pro
commandcode/gemini-3.1-flash-lite
commandcode/gemini-3.5-flash
commandcode/glm-5
commandcode/glm-5.1
commandcode/gpt-5.3-codex
commandcode/gpt-5.4
commandcode/gpt-5.4-mini
commandcode/gpt-5.5
commandcode/kimi-k2.5
commandcode/kimi-k2.6
commandcode/kimi-k2.7-code
commandcode/mimo-v2.5
commandcode/mimo-v2.5-pro
commandcode/minimax-m2.5
commandcode/minimax-m2.7
commandcode/minimax-m3
commandcode/nemotron-3-ultra
commandcode/qwen-3.6-max-preview
commandcode/qwen-3.6-plus
commandcode/qwen-3.7-max
commandcode/qwen-3.7-plus
commandcode/step-3.5-flash
commandcode/step-3.7-flash
commandcode/taste-1