Recipe Screenshot to Text
You screenshotted a recipe from a TikTok overlay, a cookbook page, or a friend's text message. Drop it here and you get a plain text file you can paste into Notes, Notion, a recipe manager, or a shared family doc. No subscription, no Honeydew or Paprika app required, and nothing leaves your browser.
Drop your recipe screenshot
or click to browse
JPG · PNG · WEBP · BMP · GIF
The five most common recipe-screenshot sources, ranked by OCR friendliness
- Cookbook page photo, in focus, good light. Near-perfect. Printed type on plain paper is what Tesseract is built for.
- TikTok / Reels caption overlay. Usually high contrast white text on dark. Reads cleanly.
- iMessage screenshot of a recipe text. Works well. Tap-and-hold copies the bubble, but if the message contains an image (someone screenshotted the recipe into the chat), you need OCR.
- Pinterest pin image with recipe text on top. Variable. Decorative fonts and color overlays trip it up.
- Handwritten card. Mostly fails. Use the cookbook-photo trick: write the recipe out in print, photograph, OCR.
How to get the cleanest result
- Zoom in before screenshotting. Small text becomes mushy when scaled. Pinch in until the text fills the screen.
- Crop out everything except the recipe. Time stamps, like buttons, and surrounding chrome confuse the layout parsing.
- Keep the screenshot rotation upright. Even a few degrees of tilt hurts accuracy.
- If it's behind a paywall blur, you can't OCR what isn't there.
From the .txt to a usable recipe
The output is plain text. From there you can paste it into:
- Notes / Apple Notes. Pinch in, paste, done.
- Notion or Obsidian. Paste into a recipe template.
- Paprika or Mela. Both let you paste raw text and auto-detect ingredients vs. instructions.
- A Google Doc shared with family. Easiest no-app option.
What about ChatGPT / AI tools?
An LLM can extract a recipe from a screenshot too, but you're handing the recipe (and any branding or personal info around it) to OpenAI's servers. If the recipe is public, fine. If it's your grandma's, or a cookbook the author asked not to be redistributed, this tool keeps the image local.
Nothing leaves the tab
The OCR engine is Tesseract compiled to WebAssembly. It runs in your browser. The "loading" you see on first use is the ~10 MB language data file caching; second use is instant.