lunaroute

LunaRoute is a high-performance local proxy for AI coding assistants like Claude Code, OpenAI Codex CLI, and OpenCode. Get complete visibility into every LLM interaction with zero-overhead passthrough, comprehensive session recording, and powerful debugging capabilities.

[{"step":"Install LunaRoute using the appropriate package manager for your system (e.g., `brew install lunaroute` for macOS or `pip install lunaroute` for Python environments).","tip":"Ensure you have Python 3.8+ installed if using the Python package. Verify installation with `lunaroute --version`."},{"step":"Configure your AI tool to route traffic through LunaRoute by setting the proxy environment variable. For OpenAI Codex CLI, use `export OPENAI_API_BASE=http://localhost:8080/v1`. For Claude Code, modify the config.json to include the proxy URL.","tip":"Use `lunaroute --init` to generate a default configuration file if you're unsure about the setup. The tool will guide you through the proxy port and logging directory selection."},{"step":"Start capturing sessions by running your AI tool as usual. LunaRoute will automatically intercept and log all LLM interactions to the specified directory (default: ~/.lunaroute/logs).","tip":"Use `lunaroute --watch` to monitor live session data in your terminal. This helps verify that the proxy is working before proceeding with analysis."},{"step":"Analyze the intercepted data using LunaRoute's built-in tools or export the logs to your preferred analysis platform. Use commands like `lunaroute --analyze --metric token-usage` or `lunaroute --export --format json` for further processing.","tip":"For large projects, filter sessions by date range or specific endpoints using `lunaroute --filter --since 2024-05-01 --endpoint /api/v1/data`. This reduces noise in your analysis."},{"step":"Implement optimizations based on the insights and monitor the impact using LunaRoute's continuous logging. Re-run analyses periodically to track improvements.","tip":"Set up automated alerts for critical metrics (e.g., latency >3s or token usage >2,000) using `lunaroute --alert --threshold latency:3s`. This ensures you catch regressions early."}]