recipe-cleanup

Recipe Cleanup

Parse the input: the first argument is either a local recipe file path or a URL to a recipe page. The optional second argument is the source URL for the original recipe. Any remaining text after the URL or file path is treated as inline recipe content (ingredients, instructions, notes).

Determine the input mode from the first argument:

• If it starts with http:// or https://, treat it as a URL — fetch the page to extract the recipe content, and treat that URL as the implicit source URL (unless a second argument overrides it)
  • YouTube watch URLs (youtube.com/watch?v=VIDEO_ID): extract the video ID and set the thumbnail to https://i.ytimg.com/vi/VIDEO_ID/hq720.jpg, then attempt to retrieve the description using the following fallback sequence:
    1. Try WebFetch on the YouTube URL first — if it returns usable description content, use that
    2. If WebFetch fails or returns no description, fall back to Playwright: navigate to the page, wait 3 seconds for it to load, then click the "...more" button to expand the full description
    3. Once the description is obtained (by either method), scan it for recipe content:
       • Recipe in description: if the description contains ingredients or instructions directly, use that as the recipe content
       • Link in description: if the description contains a link to an external recipe page (e.g. a blog post or dedicated recipe URL), fetch that URL using WebFetch and extract the recipe from there instead; record both the YouTube URL and the recipe page URL — both will appear in the output file
       • Inline text fallback: if neither is found in the description, fall back to any inline text provided in the arguments
    4. Use the channel name and video title to identify the creator
• Otherwise, treat it as a local file path — read the file