Transcribing text out of a single photo is manageable; doing it across twenty photos of scanned pages, screenshots, or handwritten notes is a genuinely tedious task nobody wants to do by hand. This tool runs optical character recognition (OCR) across an entire batch of images in one pass, reading whatever text is visible in each one and combining all of the recognised text into a single downloadable file, clearly separated by image so you can tell where each section came from.
To use it, select or drop as many images as you need — JPGs, PNGs, screenshots, or photos — into the box above, and press Extract All Text. The tool works through the images one at a time, running each through an OCR engine that runs directly inside your browser, and appends the recognised text from each image to a running document, with a clear marker showing which image each block of text came from. Once every image has been processed, the combined result is offered as a single .txt file you can download, along with a live preview so you can check it before saving.
Because every image is processed independently, you can mix very different kinds of source material in a single batch — a few scanned pages of a document alongside a couple of screenshots of a slide deck, for instance — and each one will still be read on its own terms rather than being expected to match a single layout. As with any OCR tool, image quality matters a great deal: clear, well-lit, reasonably high-resolution images of printed or clearly written text produce noticeably better results than blurry, dark, or heavily stylised source material, and neat handwriting fares better than a rushed scrawl.
This saves substantial time whenever you're digitising a stack of scanned pages, extracting text from a series of screenshots for notes or research, pulling quotes or figures out of a batch of photographed documents, or preparing a rough first-pass transcript of printed material that you'll tidy up afterwards rather than typing from scratch. Because larger batches take proportionally longer to process (each image genuinely has to be read individually), it helps to start with a smaller test batch if you're working with a very large number of images, just to confirm the quality of the results on your specific material.
For a large batch, it's worth doing a quick test run with just two or three representative images first to confirm the recognition quality on your specific material before committing to processing dozens of images at once, since that gives you a chance to retake or crop any problem photos before running the full batch.
The OCR engine runs entirely inside your browser using JavaScript — every image you add is processed locally on your own device and is never uploaded to a server. That matters particularly for scanned personal documents, notes, or anything else not meant to leave your hands. The tool is free, has no sign-up requirement, and places no limit on how many images you can process in a batch.