I checked where my AI coding tokens were going… and built a tool for it

I’m a senior SWE and use Claude Code every day. I got curious about where all those tokens were actually going, so I built optimAIzr for myself. It analyzes Claude Code + Codex usage and finds potentially wasteful patterns like huge contexts, repeated calls, expensive model usage, live recommends changes etc. It’s now at 1,500+ npm downloads in 2 weeks , and I just made it open source 🎉 It runs fully locally, reads usage/token data only (not your prompts), and needs no account. npm i -g optimaizr optimaizr profile web: optimaizr.com git: github.com/blendbunjaku/optimaizr If you try it, I’d love to know what it finds in your usage