Rescue a recipe shown in a video with a photo import
At a glance
Recipe Shelf's creator described turning recipe screenshots and handwritten cards into organized recipes using GPT-4 Vision. This workflow is listed under recipes, ocr, personal-organization.
- Agents / tools
- Recipe Shelf, GPT-4 Vision
SourceBuilding a tool for these workflows? See weekly rates.
What they did
1) Have a readable photo of a handwritten recipe card or a screenshot of the written recipe shown in a video. Open Recipe Shelf at https://recipeshelf.ai/ and use its photo import. 2) Submit the image for GPT-4 Vision to extract the recipe text, then save the resulting recipe alongside your other recipes. 3) For a TikTok source, first try pasting the link: the app attempts extraction from the caption. If the caption lacks the recipe but the video displays a written card, capture that card and use photo import instead. 4) Organize the saved item with tags or favorites so it can be found again. If a video contains no written recipe to extract, keep its embedded video as the reference rather than assuming the app reconstructed the cooking instructions.
Worth knowing
The creator explicitly distinguishes caption extraction from visual photo OCR. A saved video embed is not equivalent to an extracted, complete recipe.
