Online OCR

Online OCR accepts the pictures you are likely to have: JPG and HEIC photos from phones, PNG screenshots, WebP images saved from websites, TIFF and BMP scans, GIFs and scanned PDFs. Each file can be up to 95 MB and up to 20 go into one batch. You do not need to convert anything first; a HEIC photo from an iPhone, for example, is read directly.

The default output on this page is a plain .txt file, because the usual reason to run OCR online is to grab some text quickly: a paragraph from a book, a reference number from a letter, a quote from a slide. Plain text pastes cleanly into email, chat, notes or a search box without dragging any formatting along.

Printed pages break lines where the paper ends, not where sentences end. Keeping those breaks makes copied text awkward to reuse, so this page uses Plain text flow, which joins the lines of each paragraph into one and reads columns one after another. When the layout itself matters, for a form or a table, the other outputs below keep more of the structure.

Columns are read from left to right, top to bottom, the way a person would read a newspaper page. Headers, footers and page numbers are included as ordinary lines, so you may want to delete them from the text afterward if you are joining several pages into one document.

Recognition is only as good as the image. Text that is sharp, straight and large enough is read reliably; blurred, tilted or tiny text produces errors. Take photos square-on in good light, crop away everything that is not the text you need, and zoom in on screens before taking a screenshot so letters have more pixels.

Straighten pages and Auto rotate are on and fix small tilts and sideways photos. Enhance contrast under Advanced settings helps with faded print, pencil and receipts. Printed text is the default; for notes written by hand, switch Text type to Handwriting, or use the handwriting page, and expect to correct more.

Numbers deserve a second look whatever the quality. A misread digit in a phone number or an account reference still looks like a valid number, and no spell checker will flag it, so compare those against the original before you use them.

The classic case is text you can see but cannot select: a page in a printed book, a letter that arrived on paper, a slide shared as a picture, or a PDF that turns out to be a scan. Retyping a paragraph takes minutes and invites typos; recognizing it takes seconds, and you only have to proofread.

Short snippets are where an online tool beats installed software. Copying a serial number from a photo of a label, an address from a scanned envelope or a quote from a screenshot does not justify opening a desktop OCR program. You open this page, drop the picture in and copy the line you need from the text file.

Longer jobs work too. A scanned report of many pages can go in as one PDF, and the .txt that comes back contains every page in order. If the goal is to keep the scan but make it searchable, the OCR PDF page produces a searchable PDF instead, and the main Image to Text page offers every output with the full set of options.