Your Claude Code usage,
read in your browser.
Point this at the transcript folder Claude Code already writes on your machine. It reads the token counts, timestamps and model names out of them and builds the dashboard right here — the same page the command-line version produces.
Drop your projects folder here
Or pick it below. Nothing is sent anywhere — the reading happens in this tab.
~/.claude/projects%USERPROFILE%\.claude\projectsNothing is uploaded
There is no server here to upload to — this is a static page, and every file you choose is read by the browser itself. The tab's content-security policy blocks outbound connections outright, so you can confirm it in the network tab rather than take our word for it.
Only usage metadata is looked at: timestamps, model names, token counts, and a hashed session id. Message content, prompts, file paths and project names are never read.
Prefer the terminal?
The command-line version finds the transcripts on its own and writes a single HTML file:
npx github:heistkit/ccusage-interface
Add --serve for a live dashboard that rescans as you work. Same code, same
page — this site just hands the file-finding job to you.