Recipe Dredger

Most Mealie tooling waits for you to paste a URL. CookDex can go and find recipes for you.

Point it at a recipe site and it reads that site’s robots.txt and sitemap to enumerate candidate pages, then checks each one before importing anything. A page has to actually look like a recipe — CookDex looks for schema.org/Recipe JSON-LD with real ingredients and instructions, and cross-checks with recipe-scrapers — so listicles and how-to posts are rejected rather than imported and cleaned up later. Anything that passes is handed to Mealie’s own scraper.

It is built to be a polite crawler: robots.txt crawl-delay is honored, with a one-second floor that configuration cannot lower. Sitemaps are cached, and URLs already imported or rejected are skipped on later runs.

Preview mode reports what it would import without writing to Mealie or recording any state.

The Dredger runs as the recipe-dredger task. In the Web UI, Discover manages the sources it crawls.