Product Overview

ai-web-feeds is organized around a small public surface: the reader at /reader, the source catalog at /sources, crawlable source and topic pages, and the compact health view at /dashboard.

Main Capabilities

Curated Catalog

1. Open /reader to browse the generated article corpus, or live recent posts when the corpus has not been generated yet.
2. Switch to /sources when you need to narrow or expand the source list.
3. Return to the reader on /reader with that slice preserved in the URL.
4. Use /dashboard for catalog health and keep OPML, enrichment, and analytics management in CLI workflows.

Reading Surface

/reader is the main place to read from the catalog. It lets you:

• search the generated article library
• filter sources by topic, source type, and verification status
• open /sources, source pages, and topic pages when you need catalog detail
• keep read, saved, starred, and archived state in the browser

Export and Interop

/api/exports/opml gives you the full catalog in portable OPML formats:

• flat OPML for quick imports
• foldered OPML for readers that support grouped feeds
• filtered OPML from the current catalog slice

Optional Backend

Most of the public web app works without a Python HTTP backend. The backend is only needed for features that depend on live server-side state, including:

• recommendations
• analytics exports and summaries
• write-side search logging

Typical Workflow

1. Open /reader to read recent posts and search the article library.
2. Switch to /sources when you want to change which sources are included.
3. Use /dashboard when you want collection health.
4. Use repo commands when you want OPML files, analytics exports, or generated artifacts.