heyuehuan/dashboard-for-claude-code — explained in plain English
Analysis updated 2026-05-18
See which projects and sessions are costing the most Claude Code usage.
Check when during the day and week you actually use Claude Code the most.
Find out which tools and skills get invoked and which never do.
Spot sessions with high churn, errors, or repeated retries.
| heyuehuan/dashboard-for-claude-code | 1038lab/agnes-ai | 3eyedtiger/video2vrcemote | |
|---|---|---|---|
| Stars | 4 | 4 | 4 |
| Language | Python | Python | Python |
| Setup difficulty | easy | easy | easy |
| Complexity | 2/5 | 2/5 | 1/5 |
| Audience | developer | vibe coder | vibe coder |
Figures from each repo's GitHub metadata at analysis time.
Requires Python 3.11+ and an existing ~/.claude/projects/ directory created by Claude Code.
Dashboard for Claude Code is a local analytics tool for anyone who uses Claude Code, the terminal based coding assistant. It reads the session logs that Claude Code already saves on your own computer and turns them into a private web dashboard showing how you actually use the tool, without sending any of that data anywhere else. Once running, it answers questions a person could not easily answer by hand, such as which projects get the most use, which sessions cost the most, what time of day the tool gets used most, which tools and skills actually get invoked, and whether a person's working style leans more supervised or autonomous. It shows this through an overview page with headline numbers, a daily cost chart, a breakdown of token usage by model, a heatmap of activity by weekday and hour, a calendar style week view of sessions color coded by project, and a searchable list of individual sessions with details like duration, tokens used, tools called, and code changes. Under the hood, Claude Code writes a log file for every session on disk. This project scans those files, which it never modifies, pulls out the relevant numbers, and stores them in a small local database so that later scans only need to look at files that changed since last time. A lightweight local web server then serves the dashboard from that database. It can be started with a single command using tools like uvx or pipx, without a separate install step, or set up more permanently with the uv package manager. By default it only listens on the machine it runs on, though it can optionally be exposed to other devices on the same local network, with an optional access token for that case. It requires Python 3.11 or newer and an existing Claude Code installation, and is released under the MIT license. It is an independent community project, not made or endorsed by Anthropic.
A local, read only web dashboard that turns your own Claude Code session logs into charts of cost, token usage, tools, and activity patterns.
Mainly Python. The stack also includes Python, FastAPI, SQLite.
MIT license, use freely for any purpose including commercial use as long as you keep the copyright notice.
Setup difficulty is rated easy, with roughly 5min to a first successful run.
Mainly developer.
This repo across BitVibe Labs
double-check against the repo, no cap.