What is this?
9Router is a local AI router and token saver: it connects coding tools like Claude Code, Codex, Cursor and Cline to 40+ providers and 100+ models, auto-falls back from your subscription to cheap and free models, and compresses tool output to save 20-40% of tokens.
9Router is a local AI router and token saver: it connects coding tools like Claude Code, Codex, Cursor and Cline to 40+ providers and 100+ models, auto-falls back from your subscription to cheap and free models, and compresses tool output to save 20-40% of tokens.Is this a cloud service I sign up for, or something I run on my own machine?You run it yourself. It installs globally, serves a local web dashboard plus an OpenAI-compatible API endpoint, and your CLI coding tools send requests to that endpoint instead of to each provider directly.
You run it yourself. It installs globally, serves a local web dashboard plus an OpenAI-compatible API endpoint, and your CLI coding tools send requests to that endpoint instead of to each provider directly.How do I get it running?Install 9Router globally and a dashboard opens in your browser. Connect a free provider there (Kiro AI, or OpenCode Free with no auth), then point your coding CLI at the router. Running from source or via Docker works too.
Install 9Router globally and a dashboard opens in your browser. Connect a free provider there (Kiro AI, or OpenCode Free with no auth), then point your coding CLI at the router. Running from source or via Docker works too.Which AI coding tools can I connect to it?Claude Code, Codex, Cursor, Cline, Antigravity, Copilot, OpenCode, OpenClaw, Continue, Roo, Droid, Kilo Code, Devin CLI, Qwen Code, DeepSeek TUI and more — the README says it works with any CLI tool.
Claude Code, Codex, Cursor, Cline, Antigravity, Copilot, OpenCode, OpenClaw, Continue, Roo, Droid, Kilo Code, Devin CLI, Qwen Code, DeepSeek TUI and more — the README says it works with any CLI tool.What's the catch with the free providers?The free tiers have real limits: Kiro AI caps free usage at 50 credits/month, OpenCode Free's model list changes without notice, and iFlow, Qwen Code and Gemini CLI free tiers were discontinued in 2026. Vertex AI's $300 credit now only applies via the Vertex AI Studio endpoint.
The free tiers have real limits: Kiro AI caps free usage at 50 credits/month, OpenCode Free's model list changes without notice, and iFlow, Qwen Code and Gemini CLI free tiers were discontinued in 2026. Vertex AI's $300 credit now only applies via the Vertex AI Studio endpoint.Where can I run it — just my laptop, or a server too?Anywhere Node.js runs — macOS, Linux or Windows — and Docker images are included. Community guides also cover deploying it to Hugging Face or a VPS so it stays up around the clock.
Anywhere Node.js runs — macOS, Linux or Windows — and Docker images are included. Community guides also cover deploying it to Hugging Face or a VPS so it stays up around the clock.How does the 20-40% token saving actually work?The RTK Token Saver auto-compresses tool-result content — git diff, grep and ls output that normally burns tokens fast — cutting 20-40% of tokens per request.
The RTK Token Saver auto-compresses tool-result content — git diff, grep and ls output that normally burns tokens fast — cutting 20-40% of tokens per request.What happens when I hit a rate limit or run out of credits?9Router automatically fails over along a Subscription → Cheap → Free chain with zero downtime, can round-robin between multiple accounts per provider, and tracks your quota so you use every credit before it resets.
9Router automatically fails over along a Subscription → Cheap → Free chain with zero downtime, can round-robin between multiple accounts per provider, and tracks your quota so you use every credit before it resets.Can I connect models I host on my own machine?Yes — anything that speaks the OpenAI API shape works: llama.cpp/llama-server and vLLM for chat, whisper.cpp and faster-whisper for speech-to-text, Kokoro-FastAPI for text-to-speech.
Yes — anything that speaks the OpenAI API shape works: llama.cpp/llama-server and vLLM for chat, whisper.cpp and faster-whisper for speech-to-text, Kokoro-FastAPI for text-to-speech.Does 9Router itself cost anything?No — the router is free and MIT-licensed; you only pay whatever the underlying providers charge, and several of them offer free tiers.
No — the router is free and MIT-licensed; you only pay whatever the underlying providers charge, and several of them offer free tiers.What are similar projects?OmniRoute, casbin-gateway, opencodex, ThinkWatch-Lite, superpowers-zh, billion-context, headroom, openclaw-zero-token
OmniRoute, casbin-gateway, opencodex, ThinkWatch-Lite, superpowers-zh, billion-context, headroom, openclaw-zero-tokenWhich topics is it in?free-ai, cline, llm-proxy, rate-limiting, ai-gateway, llm-gateway, token-usage, copilot, antigravity, gemini-cli, openclaw, opencode
free-ai, cline, llm-proxy, rate-limiting, ai-gateway, llm-gateway, token-usage, copilot, antigravity, gemini-cli, openclaw, opencode