Coding.sh
Open-source, local-first observability and evals for AI agents: see what Claude Code, Codex, Cursor and others did and what it cost.
Coding.sh is an open-source, local-first observability tool for AI agents that records sessions, cost, and evaluations. It works with multiple agents including Claude Code, Codex, Cursor, OpenCode, and others. Installation is done via a single command, and it finds agents and asks before changing settings.
Session data remains on the user's machine by default, with no session data sent externally unless explicitly configured. The tool adds minimal performance overhead, as the hook writes one line to a file and exits. It is free and open source, with no additional cost beyond the user's existing model spend.
Summary drafted from the project's own website. Every sentence is backed by text on that page and was reviewed before publishing.