Upload a photo or screenshot and convert its visible text into editable content directly in your browser.
PNG, JPG or WebP, up to 20 MB.
Florence-2 can read visible text in an image using OCR. It works with screenshots, signs, labels, and photographed documents without sending your image to UtiliTools Pro servers.
It recognizes visible text in images using Florence-2.
Fix the result directly before copying it.
Send the recognized text to the clipboard.
Save content as text file.
The image remains on your device during inference.
Run the model using WebAssembly for greater compatibility.
Select a PNG, JPG or WebP image.
Click “Extract text with AI”.
Check and correct the recognized text.
Copy it or download it as TXT.
This feature uses Florence-2 Base FT in ONNX format. The model handles OCR, image descriptions, and detection through different prompts, so the browser can reuse stored model files after the first download.
Florence-2 runs through WebAssembly for broader browser compatibility. The initial download is much larger than for a typical web tool, but cached files can be reused.
Yes. Florence-2 runs locally without an API that charges per image.
It can recognize visible printed text in photos, screenshots, and documents. Accuracy depends on resolution, lighting, orientation, and font.
The first time you need the Internet to download the model files. Then the browser can reuse stored files, although offline availability depends on your cache.
No. Inference runs on your device.
It depends on the image. Florence-2 combines vision and language, but very structured documents can benefit from specialized OCR tools.
