CCR

Claude Code Router Integration Guide

This guide will help you integrate GPT-Load proxy service with Claude Code Router tool, enabling multi-model intelligent routing and code assistance features.

Prerequisites

• Ensure you have successfully deployed and started the GPT-Load service, running on http://localhost:3001 by default

• Node.js and npm environment installed

1
Tool Installation

Global Installation of Required Components

Install Claude Code

npm install -g @anthropic-ai/claude-code

Install Claude Code Router

npm install -g @musistudio/claude-code-router

Reference Documentation:Claude Code Router Official Documentation

2
Configuration File Setup

Create Configuration File

Configuration File Path

~/.claude-code-router/config.json

If the directory doesn't exist, please create the directory structure first

Configuration Example

{
  "Providers": [
    {
      "name": "gpt-load-openai",
      "api_base_url": "http://localhost:3001/proxy/openai/v1/chat/completions",
      "api_key": "sk-123456",
      "models": [
        "gpt-4.1-mini",
        "gpt-4.1-nano"
      ]
    },
    {
      "name": "gpt-load-gemini",
      "api_base_url": "http://localhost:3001/proxy/gemini/v1beta/models/",
      "api_key": "sk-123456",
      "models": [
        "gemini-2.5-pro",
        "gemini-2.5-flash"
      ],
      "transformer": {
        "use": [
          "gemini"
        ]
      }
    },
    {
      "name": "gpt-load-gemini-openai",
      "api_base_url": "http://localhost:3001/proxy/gemini/v1beta/openai/chat/completions",
      "api_key": "sk-123456",
      "models": [
        "gemini-2.5-pro",
        "gemini-2.5-flash"
      ]
    },
    {
      "name": "gpt-load-anthropic",
      "api_base_url": "http://localhost:3001/proxy/anthropic/v1/messages",
      "api_key": "sk-123456",
      "models": [
        "claude-sonnet-4-20250514",
        "claude-3-haiku-20240307"
      ],
      "transformer": {
        "use": [
          "Anthropic"
        ]
      }
    }
  ],
  "Router": {
    "default": "gpt-load-gemini,gemini-2.5-pro",
    "background": "gpt-load-gemini,gemini-2.5-flash",
    "think": "gpt-load-gemini,gemini-2.5-pro",
    "longContext": "gpt-load-gemini,gemini-2.5-pro",
    "longContextThreshold": 60000,
    "webSearch": "gpt-load-gemini,gemini-2.5-flash"
  }
}

3
Configuration Explanation

Providers Configuration Details

gpt-load-openai

OpenAI Channel Type

gpt-load-gemini

Gemini Native Format

gpt-load-gemini-openai

Gemini OpenAI Compatible Format

gpt-load-anthropic

Anthropic Claude Channel

Important Configuration Notes

• Replace localhost:3001 with your actual GPT-Load access address
• Adjust providers in Providers section according to your actual channel types configured in GPT-Load
• Path components like openai, gemini are the group names you configured in GPT-Load
• Model list should be adjusted according to the models actually supported by your channels
• In Anthropic configuration, transformer.use value should be Anthropic (capitalized)

Router Configuration

defaultDefault Model

backgroundBackground Tasks

thinkThink Mode

longContextLong Context

webSearchWeb Search

Example Configuration Note:The above example mainly configures Gemini models. You can configure other model combinations as needed, such as OpenAI or Anthropic models.

4
Launch and Usage

Launch Command

Start Claude Code Router

ccr code

After configuration is complete, use this command to start the Claude Code Router tool

Verify Configuration

Test Connection

After successful startup, try executing some basic code-related tasks to confirm that all models can respond normally. If you encounter issues, please check GPT-Load service status, configuration file syntax, and network connection.

Related Resources

Claude Code Router Official Documentation
GPT-Load Configuration Guide
Channel Types Guide