Open Source · MIT Licensed

Use Claude Code with any AI provider.

A lightweight proxy that translates Anthropic API requests into OpenAI‑compatible format. OpenAI, Gemini, Ollama — one command, zero friction.

$curl -fsSL https://ccproxy.himanshuat.com/install.sh | bash

Get Started View on npm

Terminal

# Start Claude Code with the proxy

$ ccproxy

# Start the proxy server standalone

$ ccproxy server

# View your current configuration

$ ccproxy config

How it works

Transparent by design.

Five steps, zero configuration changes to Claude Code. The proxy handles everything silently.

⚡

Intercept

Claude Code sends requests to the proxy instead of api.anthropic.com

🔄

Translate

Maps Anthropic messages, tools, and system prompts to OpenAI schema

📡

Forward

Routes the translated request to your chosen provider

⚡

Stream

Transforms provider SSE chunks into Anthropic format on-the-fly

✅

Deliver

Pipes the response back to Claude Code — fully transparent

Providers

Your model, your rules.

Switch between providers with a single flag. Each one is purpose-built for its ecosystem.

OpenAI

gpt-5.2-pro and gpt-5.2 with full streaming and tool use support.

Big modelgpt-5.2-pro

Small modelgpt-5.2

Google Gemini

Gemini 2.5 Pro and Flash with automatic top_p handling.

Big modelgemini-3.1-pro

Small modelgemini-3-flash

Ollama

Run fully local models with complete privacy. No API key required.

Big modelllama3.3

Small modelllama3.2

Quick start

Up and running in seconds.

# Run the automated installer (macOS/Linux)

$curl -fsSL https://ccproxy.himanshuat.com/install.sh | bash

# Or install globally via npm

$pnpm install -g claudebox

# Interactively configure models & providers

$ccproxy config

# Start server standalone

$ccproxy server

# Start Claude Code (auto-launches proxy)

$ccproxy

Configuration

Simple, flexible config.

Set via environment variables, CLI flags, or ~/.ccproxy/config.json

Variable	Description	Default
PREFERRED_PROVIDER	openai \| gemini \| ollama	openai
OPENAI_API_KEY	OpenAI API key	—
GEMINI_API_KEY	Google Gemini API key	—
OLLAMA_API_BASE	Ollama server URL	localhost:11434
BIG_MODEL	Model for sonnet / opus requests	provider default
SMALL_MODEL	Model for haiku requests	provider default
PORT	Server port	8082