ajmalaksar25/claude-usage-dashboard

A local, one-command dashboard that tells Claude Code users whether their Claude subscription actually saves money versus paying Claude's API list prices for the same usage. It parses your local Claude conversation store, tallies tokens/messages by model and project, prices those tokens at published per-model API rates (input/output/cache read/write), and compares the API-equivalent cost to what you’ve paid via subscription. Who uses it: Claude Code users or developers comfortable running a terminal command who want an accurate, private cost audit. Key features: - Two views: Highlights (totals, API cost, subscription paid, savings delta/multiplier, daily activity chart) and Full report (activity heatmap, model breakdown donut + table, top projects, longest conversations). - Billing sources: manual billing.json or optional Gmail invoice extraction via local OAuth (read-only). - Fully local (no data leaves your device), one-command setup (Python 3.10+), parsed-data caching for fast reloads. - Open-source (MIT) and designed to be reproducible from the README.

AI-related keywords found in this repo's description, topics, or README summary — grouped by category. Each badge links to the corresponding ranking detail page.