Why my recipe didn't parse correctly

Common causes and how to fix them — for URL, photo, and video imports.

title: "Why my recipe didn't parse correctly" description: "Common causes and how to fix them — for URL, photo, and video imports." category: "recipes" order: 2

URL imports

We parse Recipe schema.org JSON-LD first. If a site doesn't publish that, we fall back to a hand-tuned scraper for the popular ones, then a generic heuristic parser.

Common failure modes:

• Site behind a paywall — we can't bypass it.
• Recipe rendered entirely in client-side JS — we can't read post-render HTML on every site.
• Recipe lives in the comments / a forum thread — not a standard recipe URL.

Fix: open the recipe, manually edit any field after import, and send the URL via the Feedback button. We fix parsers fast.

Photo imports

Best results: a single recipe card, well-lit, no glare. Handwritten recipes parse but expect to clean up names of ingredients.

Common failure modes:

• Recipe on multiple pages — photograph each, combine in the editor.
• Glare washing out text — re-shoot in different light.
• Unfamiliar units (e.g. a regional measure) — convert manually.

Video imports

We pull the transcript via Groq Whisper, then ask Claude to extract structured ingredients + steps. The video does not need to be English — Whisper handles 50+ languages.

Common failure modes:

• No clear ingredient list (chef shows but doesn't say). Edit after import.
• Background music drowning the voice. We do better than you might expect, but not perfectly.
• Recipe + chatter mixed — feel free to delete unrelated steps.

When all else fails

Hit the floating Feedback button on web or Settings → Send feedback on mobile. Paste the URL or describe the issue. We read every one and parser improvements ship multiple times a week.

Still stuck? Email support@bragi.fit — we reply within 8 hours.