Live
OCR a PDF
Make scanned PDFs searchable
Image to Text (OCR)
LiveScreenshot, receipt, book page: extract the text from an image so you can stop retyping it. Paste the image straight from your clipboard, pick the language it is written in, and copy the text back out of the page. Line breaks come back the way they sit in the picture, and nothing is re-compressed on the way in.
What it does
- Paste a screenshot from the clipboard, or drop a file
- 13 languages, picked per image so the accents come back right
- One image a job, up to 50 MB, deleted from our server within the hour
Your files
Gone within the hour.
Files are uploaded over HTTPS, used only to run this tool, and deleted from our server automatically about an hour later. Never sold, never used to train anything.
How to convert an image to text
Add the picture
Paste a screenshot straight from your clipboard, drop a file onto the page, or click to browse. One image a job, up to 50 MB. JPEG, PNG, GIF, BMP, WebP, TIFF and the HEIC photos an iPhone takes are all accepted. A photo carrying the rotate flag your phone wrote is straightened before it is read, because sideways text is text the reader does not find at all.
Pick the language
What language is the text in? sets which of the 13 readers runs, and it defaults to English. This is the one setting worth getting right: the reader is trained per language, so a Spanish page read as English comes back with the accents wrong rather than coming back empty.
Press Read the text
The image uploads, with a progress bar you can cancel, and the reading itself usually takes a second or two. Nothing is installed and there is no account.
Copy it or download it
The text appears in the page, so Copy text puts the whole thing on your clipboard without a download. Download text.txt saves it as a UTF-8 file instead, and Read another clears the stage for the next picture.
Frequently asked questions
How do I convert an image to text for free?
Paste or drop one picture onto the page, choose the language the text is written in, and press Read the text. It runs online with no account and nothing to install. The text comes back in the page where you can copy it with one button, or download it as a plain UTF-8 .txt file. It is free and there is no watermark. The line breaks you can see in the picture come back with it: only the blank lines the reader pads the end with are trimmed, so a receipt or a book page keeps the shape it had rather than arriving as one run-on paragraph.
Which image formats can I read text out of?
JPEG, PNG, GIF, BMP, WebP, TIFF, and HEIC or HEIF, which is what an iPhone saves photos as by default. A file that is not one of those is refused at upload rather than failing later. Screenshots are the easiest case because they are already sharp: a PNG straight from the clipboard reads better than a photograph of the same screen. The reason the list is that long is that whatever you send is converted to PNG before it is read, which is also why an iPhone photo works at all: the recogniser itself has no decoder for HEIC or the more unusual TIFF flavours. PNG is the intermediate because it is lossless, so a JPEG is never re-compressed on the way in. That matters more here than it looks: JPEG artefacts collect around the edges of letters, which is exactly the detail recognition depends on.
What are the limits, and is there a daily cap?
One image per job, up to 50 MB. There is no daily cap on this tool: unlike Video to Text and Subtitle Generator, which share five speech jobs a day, reading an image only carries the site's ordinary rate limit of 20 requests a minute from one connection, so a long run of screenshots is fine. Two further limits are worth knowing because neither is about the file size. Reading is given 60 seconds, and a picture still going at that point is stopped and reported rather than left hanging: one image is a single pass, so a minute is generous, which is why this limit is far shorter than the eight minutes the PDF OCR tool gets for up to 60 pages. There is also a ceiling counted in pixels rather than bytes, so a picture with a few hundred million of them is refused even when the file itself is small, which a heavily compressed screenshot of an enormous canvas can be. Your image and the text it produced are deleted from our server automatically about an hour after the job, so save anything you need.
Which languages can it read?
Thirteen: English, Spanish, French, German, Italian, Portuguese, Dutch, Russian, Arabic, Hindi, Urdu, Chinese (Simplified) and Japanese. You pick one per image rather than the tool guessing, because the reader is trained per language and the right choice is what makes accented and non-Latin characters come back correctly. An image mixing two languages is read with the one you chose, so it is worth running twice if both halves matter.
Why does it say no readable text was found?
That message means the reader looked at the picture and got fewer than three characters out of it, so the tool tells you rather than handing you an empty file. Small, blurred, angled or low-contrast writing is the usual cause, and a sharper or straighter photo normally fixes it. It is built for printed and typed text: handwriting is not what these readers are trained on. It is worth separating from the other thing that can go wrong, because the two have different fixes: this one means the picture was read and held nothing legible, whereas a file that is damaged or in an unusual flavour of its format fails earlier with a different message and never gets as far as being read. Whatever does come back is the text alone, because the recogniser's progress chatter goes to a separate stream and never reaches the file you download.
Should I use this or OCR PDF?
Use this one when what you have is a picture and what you want is the words. Use OCR PDF when what you have is a scanned PDF and you want it to stay a PDF: it puts an invisible text layer back under the scan so the document becomes searchable and selectable, rather than handing you the text on its own. If you have a PDF that already has real text in it and you just want that text out, PDF to Text is the faster route, because it reads what the file already stores instead of looking at the pixels.