Everything You Need to Know About 9Router

9Router is a smart AI router and token saver designed for developers who use AI coding tools. It acts as a proxy between your favorite CLI tools (like Claude Code, Cursor, Codex, OpenClaw, Cline, etc.) and multiple AI providers.

It's perfect for developers who want to:

• Save 20-40% on tokens using RTK compression
• Avoid hitting rate limits with smart auto-fallback
• Maximize their subscription value with quota tracking
• Access multiple AI providers through one interface
• Use free AI models for coding without cost

9Router solves several common problems developers face with AI coding:

• Subscription waste: Quotas expire unused every month
• Rate limits: Getting stopped mid-coding when you need to focus
• Token burn: Tool outputs like git diff, grep, and ls consume tokens rapidly
• High costs: Expensive APIs ($20-50/month per provider)
• Manual switching: Constantly switching between different providers and tools

9Router provides automatic solutions for all these issues, ensuring you can code continuously without interruptions or unexpected costs.

Getting started with 9Router is simple:

1. Open the App: Navigate to 9Router Dashboard
2. Connect a Provider: Start with a FREE provider like Kiro AI or OpenCode Free (no signup needed)
3. Configure Your CLI Tool: Point your AI coding tool to 9Router's endpoint
4. Start Coding: Begin using AI models with auto-fallback and token savings

The entire process takes just a few minutes, and you can be up and running with unlimited AI coding.

9Router works seamlessly with all major AI coding tools:

• Claude Code - Anthropic's CLI tool
• OpenClaw - AI assistant for messaging apps
• Codex - OpenAI's coding assistant
• Cursor - AI-powered code editor
• Cline - Command-line AI assistant
• Antigravity - AI coding tool
• Continue - VS Code AI extension
• GitHub Copilot - AI pair programmer
• And many more...

If your tool supports custom OpenAI-compatible endpoints, it will work with 9Router.

RTK (Result Token Kompressor) is 9Router's built-in token saving feature that automatically compresses tool outputs before they're sent to the AI model.

Saves 20-40% tokens per request by compressing common outputs like:

• git diff and git status
• grep and find results
• ls and tree directory listings
• Log files and console outputs
• Large file contents

RTK works automatically - no configuration needed. It detects tool outputs and applies smart, lossless compression that preserves all the important information while dramatically reducing token count.

9Router's smart 3-tier fallback system ensures you never stop coding:

1. Tier 1 - Subscription: Uses your Claude Code, Codex, or other subscription first
2. Tier 2 - Cheap: Automatically switches to affordable providers like GLM ($0.6/1M) or MiniMax ($0.2/1M) when subscription quota is exhausted
3. Tier 3 - Free: Falls back to free providers like Kiro AI (~50 credits/month) or OpenCode Free when paid options are depleted

This happens automatically and seamlessly - you won't even notice the switch, and your coding session continues uninterrupted.

9Router supports a wide range of AI providers across three categories:

• Subscription Providers: Claude Code, Codex, GitHub Copilot, Cursor IDE, Antigravity
• Cheap API Providers: GLM, MiniMax, Kimi, OpenRouter, DeepSeek, Mistral, and 20+ more
• Free Providers: Kiro AI (~50 credits/month), OpenCode Free (no auth), Vertex AI ($300 free credits)

You can mix and match these providers in "combos" to create your perfect fallback chain.

No. 9Router itself is completely free and open-source software. It never charges you anything.

You only pay the AI providers directly:

• ✅ Subscription providers: You pay Claude Code ($20-200/mo), Codex ($20-200/mo), etc. directly on their websites
• ✅ Cheap providers: You pay GLM ($0.6/1M), MiniMax ($0.2/1M), etc. directly for usage
• ❌ 9Router: Never charges anything, ever

9Router is just a router/proxy that helps you use these services more efficiently.

Yes, the current FREE providers are genuinely free, but with some limits:

• Kiro AI: ~50 credits/month free (plus 500 trial credits for new accounts in the first 30 days). Paid tiers available above that.
• OpenCode Free: No-auth passthrough proxy. Models are auto-fetched and fluctuate over time - some may only be free during limited promos.
• Vertex AI: $300 free credits for new Google Cloud accounts (90 days). Since March 2026, use the Vertex AI Studio endpoint to consume these credits.

Discontinued free tiers: iFlow and Qwen Code free tiers were discontinued in 2026. Do not use them.

The dashboard tracks your token usage and displays estimated costs as if you were using paid APIs directly. This is not actual billing - it's a reference to show how much you're saving.

Example:

• Dashboard shows: "$290 total cost"
• Reality: You're using Kiro free models (~50 credits/month)
• Your actual cost: $0.00
• What $290 means: Amount you saved by using free models instead of paid APIs!

The cost display is a "savings tracker" to help you understand your usage patterns.

9Router takes security seriously:

• Local Processing: 9Router runs locally on your machine, not in the cloud
• API Keys: Your API keys are stored locally in your browser's local storage
• Encryption: All communications between your tools and 9Router are local
• No Data Collection: 9Router does not collect or store your code, prompts, or responses
• Self-Hosted: You have full control over where and how 9Router runs

Your data never leaves your local environment unless you specifically configure it to connect to cloud providers.

OAuth tokens are handled securely:

• Automatic Refresh: OAuth tokens are automatically refreshed before they expire
• Local Storage: Tokens are stored locally in your browser
• No Third-Party Access: 9Router never shares your tokens with third parties
• Secure Routing: All token handling happens within your local 9Router instance

You maintain full control over your authentication credentials.

We offer several support channels:

• Documentation: Comprehensive guides and tutorials available on the website
• Video Tutorials: Step-by-step video guides for setup and common use cases
• Community: Join our community for peer support and discussions
• GitHub Issues: Report bugs or request features on our GitHub repository
• Email Support: For premium users and urgent issues

Most issues can be resolved by checking the troubleshooting section or reviewing the setup guides.

If you experience issues with a specific provider:

1. Check the Dashboard: The quota tracker shows if you've hit limits
2. Try Fallback: Use a combo with multiple providers for automatic switching
3. Request Logs: Enable debug mode to see detailed request/response logs
4. Check Provider Status: Verify if the provider is having service issues
5. GitHub Issues: Report persistent issues with specific providers

The auto-fallback system usually handles provider issues automatically, but you can also create custom combos to work around specific limitations.

Yes! 9Router can be deployed on servers, VPS, or cloud platforms:

• VPS Deployment: Full control over your instance with Node.js
• Docker: Run in containers with published images available
• Cloudflare Workers: Serverless deployment on edge network
• Environment Variables: Configurable settings for production

This allows you to share your 9Router instance across multiple devices or team members while maintaining security and control.

Custom combos let you create your perfect fallback chain:

1. Go to Dashboard → Combos → Create New
2. Name your combo (e.g., "premium-coding")
3. Add models in priority order:
   • 1. Subscription model (e.g., cc/claude-opus-4-7)
   • 2. Cheap backup (e.g., glm/glm-5.1)
   • 3. Free fallback (e.g., kr/claude-sonnet-4.5)
4. Save and use the combo name in your CLI tool

Example: "premium-coding" will automatically try your subscription first, then cheap options, then free models if needed.