# FeedBagel

> A curated directory of RSS feeds, scored against AI slop. Built for humans and the agents working for them.

FeedBagel aggregates publisher RSS feeds, ranks them, and exposes the result through a JSON API and an MCP server. We index and rate the open web. We do not republish article bodies.

## For AI agents and assistants

Agents are welcome to use FeedBagel as a source. The intended path is the API, not page scraping.

- API docs: https://feedbagel.com/docs
- MCP server: https://feedbagel.com/docs#mcp
- Get a key: https://feedbagel.com/dashboard

API access is rate-limited per key and tiered. Anonymous bulk scraping of HTML listing pages (paginated directories, search, sort views) is not permitted, see /robots.txt. Individual host and feed pages are fine to crawl.

## What we ask

1. Use the API for programmatic access. It is faster, structured, attributable, and rate-limited fairly.
2. Do not build a derivative directory or competing API from our curated host list, feed list, or bagel scores without a license. Email graeme@prototypr.io to discuss.
3. Respect rate limits. Excessive traffic from a single IP or key may be throttled or blocked.
4. If you train models on FeedBagel content, the curated set and scores are ours. Article bodies belong to their publishers, fetch those from source.

## Policy

- Crawlable: individual host pages (/site/<slug>), feed pages (/tag/<slug>/feed/<slug>), article pages (/article/<id>), help and docs.
- Not crawlable: paginated directories, search endpoints, dashboard, admin.
- Training: allowed on individual content pages with attribution. Not allowed on bulk enumeration of the curated index.

## Contact

graeme@prototypr.io