Show HN: Open-source CLI to see your AI coding token usage and compare it (github.com)

A new open-source command-line tool, "whoburnedmore," has been announced, enabling users to monitor their AI token usage from Claude Code sessions on their local machine. This fast, 100% local CLI reads session transcripts without sending any data to external servers, ensuring complete privacy. It allows developers to see their token consumption in both token counts and estimated costs, broken down by AI model and project. With insights into token usage over the last 30 days and a user-friendly HTML dashboard, the tool links raw data to financial implications, helping developers make informed decisions about their AI usage. Significant for the AI/ML community, this tool addresses the growing concern over AI operational costs and resource management. By providing a transparent breakdown of token usage by project and model, developers can identify where optimizations are needed. Additionally, the tool's design emphasizes security and usability—requiring no account setup, and making no network requests, it preserves users' data integrity. The ability to access detailed reports directly on one's device fosters a greater understanding of AI cost structures, thereby helping to refine budgeting and usage strategies in AI project development.