Self-hosted GitHub profile cards showing AI coding usage across tools, with per-tool costs and a 26-week activity view. Runs locally with daily SVG regeneration and no external services.
Collecting history — the radar snapshots this repo daily. The trend line appears after 3 days of data (1 so far).
What it is
A self-hosted GitHub profile card that shows how many tokens you've burned across AI coding tools, with API-equivalent costs in multiple currencies, per-tool splits, and a 26-week activity heatmap. It runs locally against your own logs and commits an SVG to your profile repository.
How it works
ccusage (each device) ──► cards/devices/<device>.json ──► merged SVG ──► profile README ▲ ▲ ▲ local CLI logs overwritten per run all devices combined
A scheduler runs the script once a day. Each run recomputes everything from local logs, fetches FX rates, and commits the regenerated SVG.
Getting started
Requirements: Node 18+, GitHub CLI logged in (gh auth login, repo scope), and a profile repo named after your username.
- Fork and clone:
git clone https://github.com/YOURNAME/ai-coding-usage-card.git
cd ai-coding-usage-card
-
Choose a permanent device ID.
-
Test run:
USAGE_CARD_REPO=YOURNAME/YOURNAME USAGE_CARD_DEVICE=macbook-work node usage-card.mjs
# [2026-07-16T...] 1 device(s), 5 cards updated @ ab12cd3: 12.6B tokens | $13,190 | Claude Code $13,190 | Codex $0.04
-
Embed in your profile README using the provided image tags.
-
Schedule daily via Windows Task Scheduler or macOS/Linux cron as shown in the README.
Recent releases
- v1.0.0 — first stable release (2026-07-16): First stable release.
Traction
18 stars, 6 forks, 0 open issues (as of latest facts).
License
MIT