Analysis updated 2026-05-18
Generate a weekly Markdown report of how your Claude Code time was spent.
See a category breakdown of coding, writing, and research activity from your sessions.
Get an AI-written narrative summary of what you shipped each week.
Export your usage report to Obsidian for personal knowledge tracking.
| atomchung/ccstory | 0whitedev/detranspiler | 2951461586/mulerun-pool | |
|---|---|---|---|
| Stars | 21 | 21 | 21 |
| Language | Python | Python | Python |
| Setup difficulty | easy | hard | moderate |
| Complexity | 2/5 | 4/5 | 3/5 |
| Audience | developer | developer | developer |
Figures from each repo's GitHub metadata at analysis time.
Runs entirely on local Claude Code session logs, optional narrative mode uses your existing Claude Code session.
ccstory is a command-line tool for people who use Claude Code, Anthropic's AI coding assistant, and want a human-readable summary of how they've been spending their time with it. Where ccusage (a companion tool) tells you what your Claude API bill was, ccstory reads the same local log files and tells you what the tokens actually went toward: coding, writing, research, or whatever categories you define. It reads Claude Code's session logs stored in ~/.claude/projects/ on your machine, groups the sessions into time buckets and activity categories based on folder name rules or by having Claude analyze the content of each session, and produces a formatted report showing active hours, number of sessions, token output, cost, and a breakdown of time spent per category. An optional narrative mode calls Claude locally to write 2-3 sentence summaries per category and one-liner descriptions per session, turning the numbers into a written account of what you actually shipped that week. The full report is saved as a Markdown file, an Obsidian export option adds YAML frontmatter for use in personal knowledge management vaults. You would use this for weekly retrospectives, status updates, or just to understand where your AI usage is going after you see a surprising bill. Everything runs locally, no telemetry, no external service, and the optional AI narrative feature uses your existing Claude Code session rather than a separate API key. The tech stack is Python 3.10 or later, installed as a package via pipx. The full README is longer than what was provided.
A command-line tool that reads your local Claude Code session logs and turns them into a readable weekly report of what you actually worked on.
Mainly Python. The stack also includes Python, pipx.
Setup difficulty is rated easy, with roughly 5min to a first successful run.
Mainly developer.
This repo across BitVibe Labs
Verify against the repo before relying on details.