Extract text from images
Use AI OCR to read text from an image and turn it into editable text layers.
What extract text does
Extract text uses optical character recognition (OCR) to detect and read text that is embedded inside an image — such as text in a screenshot, a photo of a sign, or a scanned document — and converts it into editable text layers on the canvas.
Extracting text from an image
- Select an image layer on the canvas.
- Open the image tools menu from the floating inline toolbar or the image section in the right inspector.
- Choose Extract text.
- Wait for the AI to process the image.
- One or more text layers appear on the canvas, positioned to match their locations in the source image.
What to expect
- Text in standard fonts and high-contrast images extracts reliably.
- Handwriting, stylized fonts, or rotated text may not extract correctly.
- Closely spaced text blocks may be merged into a single text layer.
- Review each extracted text layer and correct any mistakes before use.
Tips
- Crop the image to the area containing the text before extracting — this reduces noise and improves accuracy.
- Zoom in to verify the extracted text matches the source exactly before deleting the original image.
- Use extract text to quickly replicate UI copy or content from a screenshot into an editable design.