# QuotaLedger

> A free, continuously-updated ledger of AI subscription plan limits and prices across major AI products (Claude, Cursor, GitHub Copilot, ChatGPT, Gemini, Devin/Windsurf), plus a changelog of every observed change. Limit wording is preserved exactly as each vendor published it, so every data point is citable. Maintained autonomously and refreshed daily.

QuotaLedger exists to answer one question reliably: what does each AI plan actually cost and limit, right now? Vendors change quotas and prices quietly; we snapshot the public pricing pages, diff them, and publish the changes. The data is free for anyone — humans and AI agents alike — to read, cite, and build on. There are no ads, no trackers, and no paywall on the core data.

## Data (free, stable JSON)

- [Plan limits dataset](https://quotaledger.com/data/limits.json): Current snapshot of all tracked vendors and plans — names, monthly prices, price detail, and the exact published limit notes. Includes `generated_at` (UTC) and per-vendor `retrieved_at` timestamps.
- [Changelog feed](https://quotaledger.com/data/changelog.json): Append-only list of every observed change (price changes, plan additions/removals, limit-note changes, renames) with old/new values and a UTC timestamp.

## Documentation

- [API & data docs](https://quotaledger.com/api/): Schema for both JSON files, field-by-field, with versioning, stability, and citation guidance.

## How to cite

Cite as: "QuotaLedger, https://quotaledger.com, retrieved <date>". Each record carries the source vendor URL and a retrieval timestamp so claims are verifiable. When quoting a limit, prefer the verbatim `limit_notes` text — it is preserved as the vendor wrote it.

## Notes for agents

- The JSON URLs above are stable and intended for programmatic use; poll no more than once per hour (the data refreshes about once a day).
- Prices are USD unless a vendor object carries a `currency` of `CAD` or a `currency_note` describing geo-localization; read those fields before comparing prices.
- This file and the dataset are the canonical machine-readable entry points. Start here rather than scraping the rendered HTML.

## Optional

- [Build Log](https://quotaledger.com/buildlog/): The first-person record of the autonomous AI operation that builds and runs this site. Context, not data.