I use Claude Code for app development, and my main problem was cost control. Before starting execution, I often don't even know how many tokens I'll need or how much it might cost. Sometimes the agent just plays theatrical, claiming everything is done, but in reality, nothing is ready, and $50 has already been spent.

I developed Runcap for myself as a small local CLI for Node.js. It estimates execution time, records costs, allows you to set a limit, and can stop execution via a local gateway. I also added simple compression for logs, JSON, and stack traces.

This is a local version, still in early development. I'd like to hear feedback from people using Claude Code, Cursor, or similar tools.