Live demo of Tokdash — a local token & cost dashboard for AI coding tools (Codex, Claude Code, OpenCode, Gemini CLI, OpenClaw, Kimi CLI, pi-agent, GitHub Copilot CLI, Hermes).
The data on this page is fully synthetic. A small in-browser shim (
static/mock-api.js) intercepts every/api/*request and returns deterministic sample data, so the unmodified Tokdash frontend can run as a static site. Nothing is uploaded; nothing is read from your machine.
- Switch tabs (Overview / Sessions / Heatmap / Pricing).
- Change the period or pick a custom date range.
- Toggle light/dark and the 10 style themes.
- Click into a synthetic Codex / Claude / OpenCode session.
- On the Stats tab, switch the Daily Activity metric chip to Energy to recolor the heatmap by estimated energy per day, and read the new Total Energy (kWh) row in the Month Stats sidebar (estimated entirely in the browser from token counts × model-family
J/tokencoefficients; order-of-magnitude only). - Browse the read-only demo pricing snapshot.
python3 -m http.server 8000
# open http://localhost:8000| File | Purpose |
|---|---|
index.html |
Tokdash dashboard shell with demo banner and synthetic backend wiring. |
static/themes.css |
Verbatim copy of the production stylesheet. |
static/theme-config.js |
Verbatim copy of the production theme palettes. |
static/mock-api.js |
Demo-only fetch shim that builds and serves synthetic data. |
pricing_db.json |
Sanitized pricing snapshot for the read-only Pricing tab. |
To refresh the demo when Tokdash itself changes, re-copy the upstream static assets,
keep the demo-only edits in index.html, and sanitize pricing_db.json so the public
demo does not advertise unreleased or placeholder model ids.
MIT — see LICENSE.