llm-cli-proxy

Use your Claude, Gemini, or ChatGPT subscription from any OpenAI-compatible tool — no API keys needed.

Wraps the official Claude Code, Gemini CLI, and Codex CLIs and exposes a local OpenAI-compatible HTTP API. Point any tool that supports OPENAI_BASE_URL or ANTHROPIC_BASE_URL at it and your subscription handles the billing.

npx llm-cli-proxy --provider claude --port 3456

Prerequisites

Node.js 20+ and at least one CLI installed and authenticated:

Provider	Subscription	Install	Authenticate
Claude	Claude Pro or Max	npm install -g @anthropic-ai/claude-code	Run claude and sign in
Gemini	Gemini Advanced (Google One AI Premium)	npm install -g @google/gemini-cli	Run gemini and sign in
Codex	ChatGPT Plus, Pro, or Team	npm install -g @openai/codex	Run codex and sign in

Installation

npm install -g llm-cli-proxy

Or use without installing:

npx llm-cli-proxy --provider claude

Usage

Start one instance per provider you want to use:

llm-cli-proxy --provider claude   # listens on port 3456
llm-cli-proxy --provider gemini   # listens on port 3457
llm-cli-proxy --provider codex    # listens on port 3458

Then point your tool at the local URL:

# For tools using the Anthropic SDK
export ANTHROPIC_BASE_URL=http://localhost:3456/v1

# For tools using the Google Gemini SDK
export GOOGLE_GEMINI_BASE_URL=http://localhost:3457/v1

# For tools using the OpenAI SDK
export OPENAI_BASE_URL=http://localhost:3458/v1

Some SDKs will refuse to initialize without an API key env var set. If you hit that, set it to any non-empty string — the proxy ignores it.

Options

llm-cli-proxy --provider <claude|gemini|codex> [options]

  --provider, -p  <name>   Provider: claude, gemini, or codex (required)
  --port          <port>   HTTP port (default: 3456/claude, 3457/gemini, 3458/codex)
  --workspace, -w <dir>    Working directory for the CLI (default: current directory)
  --model, -m     <model>  Model override
  --version, -V            Show version
  --help, -h               Show help

API

POST /v1/chat/completions

# Non-streaming
curl http://localhost:3456/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{"model":"claude-sonnet-4-6","messages":[{"role":"user","content":"Hello"}]}'

# Streaming
curl http://localhost:3456/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{"model":"claude-sonnet-4-6","messages":[{"role":"user","content":"Hello"}],"stream":true}'

GET /v1/models

Returns available models for the running provider.

GET /health

{
  "status": "ok",
  "provider": "Claude Code",
  "session_active": true,
  "model": "claude-sonnet-4-6",
  "workspace": "/your/project"
}

Models

Provider	Default	Available
Claude	claude-sonnet-4-6	claude-sonnet-4-6, claude-opus-4-6
Gemini	gemini-2.5-flash	gemini-2.5-flash, gemini-2.5-pro
Codex	codex-mini-latest	codex-mini-latest

How It Works

• Claude & Gemini — spawns the official CLI per request using --resume to maintain conversation context. The CLI handles authentication automatically.
• Codex — calls the Codex API endpoint directly using the OAuth token stored by the Codex CLI. This avoids a known bug in exec resume that causes the system prompt to be re-injected on every turn, multiplying token usage.

License

MIT