AI Gateway for Subscription-Based LLMs
Turn your Claude Pro, GitHub Copilot, and Gemini subscriptions into standard LLM APIs. No API keys needed.
- Multi-Provider — Claude, Copilot, Gemini, Codex, Qwen, Kiro, and more
- Multi-Format — OpenAI, Anthropic, Gemini, Ollama compatible endpoints
- Multi-Account — Load balance across accounts, auto-retry on quota limits
- Zero Config — OAuth login, no API keys required
- Protocol Translation — IR-based translator converts between all formats
# macOS / Linux
curl -fsSL https://raw.githubusercontent.com/nghyane/llm-mux/main/install.sh | bash
# Windows (PowerShell)
irm https://raw.githubusercontent.com/nghyane/llm-mux/main/install.ps1 | iex# 1. Login to a provider
llm-mux --antigravity-login # Google Gemini
llm-mux --claude-login # Claude Pro/Max
llm-mux --copilot-login # GitHub Copilot
# 2. Start the server (runs as background service after install)
llm-mux
# 3. Verify
curl http://localhost:8317/v1/models| Format | Endpoint | Use With |
|---|---|---|
| OpenAI | /v1/chat/completions |
Cursor, Aider, LangChain |
| Anthropic | /v1/messages |
Claude Code, Cline |
| Gemini | /v1beta/models/{model}:generateContent |
Gemini CLI |
| Ollama | /api/chat |
Open WebUI |
Base URL: http://localhost:8317
API Key: unused
OpenAI format:
curl http://localhost:8317/v1/chat/completions \
-H "Content-Type: application/json" \
-d '{"model": "gemini-2.5-pro", "messages": [{"role": "user", "content": "Hello!"}]}'Anthropic format:
curl http://localhost:8317/v1/messages \
-H "Content-Type: application/json" \
-H "anthropic-version: 2023-06-01" \
-d '{"model": "claude-sonnet-4", "max_tokens": 1024, "messages": [{"role": "user", "content": "Hello!"}]}'See API Reference for all endpoints.
| Guide | Description |
|---|---|
| Installation | Install options, update, uninstall |
| Providers | All supported providers and login commands |
| Configuration | Config file reference |
| API Reference | Supported API formats and endpoints |
| Integrations | Cursor, VS Code, Aider, LangChain, etc. |
| Docker | Container deployment |
| Service Management | Background service on macOS/Linux/Windows |
| Troubleshooting | Common issues and solutions |
| Provider | Login Command | Models |
|---|---|---|
| Google Gemini | --antigravity-login |
gemini-2.5-pro, gemini-2.5-flash |
| Claude | --claude-login |
claude-sonnet-4, claude-opus-4 |
| GitHub Copilot | --copilot-login |
gpt-4.1, gpt-4o, gpt-5 |
| OpenAI Codex | --codex-login |
gpt-5 series |
| Qwen | --qwen-login |
qwen-coder |
| Kiro | --kiro-login |
Amazon Q models |
| Cline | --cline-login |
Cline API |
| iFlow | --iflow-login |
iFlow models |
See all providers for details and model lists.
Works with any OpenAI/Anthropic/Gemini-compatible tool:
Editors: Cursor, VS Code (Cline, Continue), Zed
CLI: OpenCode, Claude Code, Codex CLI, Gemini CLI
Frameworks: LangChain, LlamaIndex, Vercel AI SDK
Web UIs: Open WebUI, LibreChat
See integration guides for setup instructions.
flowchart LR
subgraph Input["API Formats"]
A[OpenAI]
B[Anthropic]
C[Gemini]
D[Ollama]
end
subgraph Core["llm-mux"]
E[IR Translator]
end
subgraph Output["Providers"]
F[Gemini]
G[Claude]
H[Copilot]
I[Codex]
J[...]
end
A --> E
B --> E
C --> E
D --> E
E --> F
E --> G
E --> H
E --> I
E --> J
MIT License — see LICENSE