Claude Code Proxy

Claude Code Proxy (Go)

A lightweight HTTP proxy that enables Claude Code to work with OpenAI-compatible API providers including OpenRouter (200+ models), OpenAI Direct (GPT-5 reasoning), and Ollama (free local inference).
⚠️ Early Stage / Beta Software
This project is in active development and has limited production testing. While core functionality works, edge cases and some provider-specific features may have issues. Use with caution in production environments.
Feedback welcome! Please report issues at https://github.com/nielspeter/claude-code-proxy/issues

Features

- ✅ Full Claude Code Compatibility - Complete support for all Claude Code features
  - Tool calling (read, write, edit, glob, grep, bash, etc.)
  - Extended thinking blocks with proper hiding/showing
  - Streaming responses with real-time token tracking
  - Proper SSE event formatting
- ✅ Multiple Provider Support - OpenRouter, OpenAI Direct, and Ollama
  - OpenRouter: 200+ models (GPT, Grok, Gemini, etc.) through single API
  - OpenAI Direct: Native GPT-5 reasoning model support
  - Ollama: Free local inference with DeepSeek-R1, Llama3, Qwen, etc.
- ✅ Adaptive Per-Model Detection - Zero-config provider compatibility
  - Automatically learns which parameters each model supports
  - No hardcoded model patterns - works with any future model/provider
  - Per-model capability caching for instant subsequent requests
- ✅ Pattern-based routing - Auto-detects Claude models and routes to appropriate backend models
- ✅ Zero dependencies - Single ~10MB binary, no runtime needed
- ✅ Daemon mode - Runs in background, serves multiple Claude Code sessions
- ✅ Fast startup - < 10ms cold start
- ✅ Config flexibility - Loads from ~/.claude/proxy.env or .env
- ✅ Passthrough mode - Optional direct proxying to Anthropic API…