Image to text, on your own PC

Windows reads it, on your computer

Windows 10 and 11 come with their own text recognition (OCR), the kind that turns a picture of words into words you can copy. Octoolo hands it your picture and shows what it finds. There is nothing extra to download, no account with an OCR service, and no internet needed: the picture is read where it is, on your PC.

That matters for the pictures people most often copy text from: a photo of a letter, a scanned bill, a screenshot of a chat, a picture of an ID card or a whiteboard after a meeting. None of it goes to a website, so none of it can be kept by one.

The picture can be a JPG, PNG, WebP, BMP, GIF or AVIF file, or a screenshot pasted straight in. One picture is read at a time; Open another clears the window for the next one.

Which languages it reads

Windows reads the languages it has text recognition for, and the Language list shows exactly which ones are on your PC. Your Windows languages uses the languages in your Windows settings and suits most pictures.

Picking the right language matters most for letters with accents and marks, such as é, ñ, ü or ç: read with the wrong language chosen, they can come out wrong or go missing. The text pane's title names the language that was used.

To read another language, add it to Windows: Settings, Time & language, Language & region, then Add a language (on Windows 10 the page is called Language). For languages Windows can recognize, it then shows up in Octoolo's list. If a PC has no text recognition at all, Octoolo says so and points to the same place in Settings.

What comes out, and what to expect

The result is plain text, one line of the picture per line of text, with the number of lines shown above the picture. Each word that was found is marked on the picture, so you can see at a glance what was missed.

• Clear print reads best: screenshots, documents, signs and labels.
• Layout is not kept. Columns, tables, bold and font sizes come out as plain lines, and a table becomes rows of words to tidy by hand.
• Handwriting, curved or tiny text and busy backgrounds read poorly or not at all.
• Very large pictures are scaled down to what Windows reads in one go, so small print in a huge photo can be lost. Crop to the part you need first.

To turn the lines into flowing paragraphs, paste them into remove line breaks in Text Studio. If nothing is found, Octoolo says “No text was found in this picture.”

Screenshots, photos of pages and PDFs

Screenshots: press Win Shift S, select an area, then press Ctrl V in Octoolo's reading window. No file is needed.

Photos of a page or a whiteboard: straighten and crop it in the Photo Editor first (rotate, crop), add Brightness/Contrast from Image › Adjustments if the writing is faint, export it, then read the exported copy. Level, sharp lines of text read better than tilted or blurry ones.

Receipts and labels: photograph them flat, filling the frame, in even light. A shadow across the print is a common reason for missed words.

PDFs: Image to text reads pictures, not PDF files. For a scanned PDF, open it in the PDF Editor and save its pages as pictures with Save the pages as pictures… (see PDF to JPG), then read the page you need here.