How to get text out of a picture on your PC
Drop or paste a picture into Read by Zaweba and it reads the words, so you can copy them or save them as a text file, all offline on Windows.
The quickest way is to open the picture in Read by Zaweba, a free text recognition app for Windows, and copy the text it finds. You can drop the file on the window or paste it with Ctrl+V, and Read reads the words on your own PC. From there you copy everything to the clipboard or save it as a plain text file.
Pictures with words in them turn up everywhere: a photo of a notice board, a screenshot of an error message, a scan a colleague emailed, a page of a recipe book. Typing it all out again is slow and full of mistakes. Text recognition, usually called OCR, does the typing for you.
What you need
A Windows PC with Read by Zaweba installed, and the picture. Read takes PNG, JPEG, TIFF, BMP, GIF, WEBP and HEIC files, using the image codecs that Windows already has. WEBP and HEIC need extensions from the Microsoft Store, and if Windows can't decode one of those files Read tells you exactly which extension to install rather than failing quietly.
There's no account and nothing to sign in to. The picture never leaves your computer, which is a relief when it's a payslip, a letter from the bank or anything else you wouldn't paste into a website.
Step by step
Open Read. Drag the picture from File Explorer onto the window, or copy it and press Ctrl+V. Paste works for a picture sitting on the clipboard (say, one you copied from a browser) and for files you copied in Explorer. You can also open a whole folder if you have a pile of them.
If the picture is a phone photo, Read turns it the right way up from the orientation the phone stored with it, so a photo taken sideways is read the right way round. For a photo of a page there is more clean-up on offer, such as finding the page and flattening it, and the Document tab tells you what was done.
Read then picks an engine. For a single quick picture it usually uses Windows' own OCR when Windows has the language; for columns, tables and anything bigger it uses the bundled Tesseract engine. The engine it chose, and why, is shown above the text.
When the words appear, look them over. The picture and the text sit side by side, and you can edit the text directly. Then copy all the text to the clipboard and paste it wherever it's going, or save it as a .txt file. Read saves text files as UTF-8 with a byte order mark, so accented letters and symbols survive when you open them in other programs.
Getting a clean result
The sharper the picture, the better the reading. Text that is tiny in the picture or badly blurred gives the engine little to work with, and the Help tab is honest that handwriting, low resolution pictures and decorative fonts are unreliable. If you can take the photo again, fill the frame with the text and keep the phone steady.
When Tesseract does the reading, any word it's less than 70 percent sure of is shaded amber, both on the picture and in the text. Press F8 to jump to the next one and fix it before you copy. Windows OCR doesn't give confidences, so if you want those checks, choose Read again with Tesseract from the Read menu.
When the text is on your screen instead
Sometimes the words aren't in a file at all; they're in a window, a video frame or a program that won't let you select text. For that, Read has a screen text grab. Press the hotkey (Win+Shift+T unless something else already uses it), drag round the text, and it's copied straight to the clipboard with a short notice. It's the same reading, without saving a picture first.
Either way, the result is ordinary text you can edit, search and reuse, made on your own PC by Read by Zaweba.
