Quick answer: CookBook’s importer now combines AI vision + structured-data parsing, so it can read most recipe websites, even when they skip industry-standard markup. If parts of the recipe are still missing, the page may be image-only or locked behind paywalls/scripts our AI can’t reach. You can always copy-paste the text or scan a screenshot.
Why imports sometimes fail - even with AI
Text locked inside images
When all the ingredients and instructions are baked into a JPG or Canva collage, there’s no actual text for the AI to read.
Dynamic or pay-walled pages
Some blogs load the recipe only after you scroll, click a button, or log in. CookBook’s crawler sees the page before that content appears, so it can’t extract the data.
Ingredients in custom tables
A few sites split the ingredient list across complex table cells or widgets. The AI may merge lines, drop quantities, or get units wrong because the layout isn’t a standard list.
Social posts with minimal captions
A TikTok or Instagram Reel that just says “full recipe in bio” gives the importer nothing to work with beyond a video link. If the linked site also lacks structured data, the import can fail.
Unusual narrative formatting
Some bloggers weave ingredients and steps into long paragraphs without clear headings or line breaks. The AI struggles to decide where the ingredient list ends and the instructions begin.
What you can do
Try the Print view
Many sites have a Print Recipe button that delivers cleaner HTML.
Copy-paste fallback
Copy the ingredient list + steps → open CookBook → Paste → Confirm → Save.
Use the Recipe Scanner
Screenshot the recipe (or photo from a laptop screen) → Add (+) → Scan recipe image and let AI recipe scanner turn it into text.
Send us the link
Tap Submit a request with the URL and a screenshot of the missing data. We feed tricky pages back into the training set.