Choose a picture and Photopip reads its text with an on-device OCR model, then lists every recognized line in the card above the canvas. Click Copy text to put it on your clipboard. Printed Chinese and English work well; handwriting and decorative lettering mostly don't.

1. Click Choose an image, or drop a screenshot, scan or photo of a document on the page.
2. Wait while the text is recognized. The first run downloads two models, about 16 MB together; after that recognition runs on your device, a few seconds per image.
3. Read the lines in the card. Each detected line is on its own row, in reading order; lines the model is unsure about are left out rather than guessed.
4. Click Copy text and paste the text where you need it. To erase the text from the picture instead, use Remove text from image.

Can you read the text in this picture for me?

Yes, and here the reading happens on your device: a PP-OCRv4 model with a 6623-character set covers Chinese, English, digits and common punctuation. You get plain text, one line per row, with the spaces between English words restored by measuring the gaps in the image. Formatting isn't kept, so a table comes out as a list of lines.

Why was some of the text missed?

Detection runs at 1280 pixels on the long side, so small type in a large screenshot can be skipped: click Keep editing, crop to the paragraph with the Crop Tool (C) and run it again. Lines recognized with a confidence under 0.5 are dropped. Handwriting, calligraphy, vertical text and signs photographed at a steep angle are the usual failures.

Is the image uploaded?

No. The two models (text detection 4.7 MB, recognition 10.9 MB) download to your browser once, and recognition runs in a worker on your device. If nothing is found, the card says No text recognized in the image rather than inventing any.

Does it work on PDFs?

Yes, in two steps: open the page on the PDF to JPG page first (300 ppi keeps small print legible), save the JPG, then come back here and read it.

Updated 2026-10-21

More toolsVideo to GIFAVIF to JPGSVG to PNGPDF to JPGPDF to PNGRemove peopleProduct photo on whiteTransparent background

Choose a picture and Photopip reads its text with an on-device OCR model, then lists every recognized line in the card above the canvas. Click Copy text to put it on your clipboard. Printed Chinese and English work well; handwriting and decorative lettering mostly don't.

1. Click Choose an image, or drop a screenshot, scan or photo of a document on the page.
2. Wait while the text is recognized. The first run downloads two models, about 16 MB together; after that recognition runs on your device, a few seconds per image.
3. Read the lines in the card. Each detected line is on its own row, in reading order; lines the model is unsure about are left out rather than guessed.
4. Click Copy text and paste the text where you need it. To erase the text from the picture instead, use Remove text from image.

Can you read the text in this picture for me?

Yes, and here the reading happens on your device: a PP-OCRv4 model with a 6623-character set covers Chinese, English, digits and common punctuation. You get plain text, one line per row, with the spaces between English words restored by measuring the gaps in the image. Formatting isn't kept, so a table comes out as a list of lines.

Why was some of the text missed?

Detection runs at 1280 pixels on the long side, so small type in a large screenshot can be skipped: click Keep editing, crop to the paragraph with the Crop Tool (C) and run it again. Lines recognized with a confidence under 0.5 are dropped. Handwriting, calligraphy, vertical text and signs photographed at a steep angle are the usual failures.

Is the image uploaded?

No. The two models (text detection 4.7 MB, recognition 10.9 MB) download to your browser once, and recognition runs in a worker on your device. If nothing is found, the card says No text recognized in the image rather than inventing any.

Does it work on PDFs?

Yes, in two steps: open the page on the PDF to JPG page first (300 ppi keeps small print legible), save the JPG, then come back here and read it.

Updated 2026-10-21

More toolsVideo to GIFAVIF to JPGSVG to PNGPDF to JPGPDF to PNGRemove peopleProduct photo on whiteTransparent background