OnlineOCR
Pulls the text out of a scan or photo and hands it back as a Word or Excel file laid out like the original, not a wall of plain text.
Overview
OnlineOCR reads the words in a picture. You give it a scan, a photo of a page or a PDF that is images rather than text, and it hands back an editable Word, Excel or text file. It states the output keeps the original layout. A scanned table returns as Excel rows and columns, not a run-on line. It reads 46 languages and takes multipage PDF and TIFF.
It runs in a browser. Guests get 5 files an hour and 15 MB per file, with no account. Registering is free and adds 50 pages, 200 MB uploads and PDF or RTF output. Paid plans sell pages in bulk, by subscription or as a one-time lifetime package. The vendor states guest uploads are deleted after conversion, and registered users' files are kept for a month.
Key features
When to use
Use OnlineOCR if:
- You have a scanned document or a photo of a page and you want the text as a file you can change.
- Your page has a table in it. The Excel output returns it as rows and columns rather than plain text.
- The layout matters. It states the output keeps the same arrangement as the original rather than dumping the words in a stream.
- Your document is not in English. It handles 46 recognition languages.
- You have a long scanned PDF or a multipage TIFF. It takes both, and registering lets you convert only the pages you need.
- You want one page done now, with no account and nothing installed.
When NOT to use
Skip OnlineOCR if any of these describe you:
- You want the words and nothing else. A simpler image to text tool is quicker when layout and file formats do not matter.
- You have a lot to get through. Guests get 5 files an hour and 15 MB per file, so volume means registering or a paid plan. Compare other image to text tools.
- Your document must not sit on someone's server. Registered users' files are stored for a month, and the work happens in the cloud.
- Your source is handwriting. Printed text is what OCR handles best, and Pen2txt is built for handwritten notes.
- You expect a perfect transcript. OCR misreads characters, especially on a poor scan, so you have to proofread the result.
Frequently asked questions
You can convert files as a guest without an account, at 5 files an hour and 15 MB per file. Registering is also free and gives you 50 free pages and uploads up to 200 MB. It also adds PDF and RTF output, ZIP archives, and the option to convert a page range instead of a whole PDF. Beyond that, paid plans sell pages in bulk, as a subscription or a one-time lifetime package.
It depends on the account. The vendor states that documents uploaded under the free guest account are deleted automatically after conversion, while files belonging to registered users are stored for a month. If the document is sensitive, guest mode keeps it on their servers for the shortest time.
Yes. It exports to Excel and states the output keeps the same layout as the original, so a table of rows and columns comes back as cells rather than a paragraph of numbers. Check the result against the scan, because a misread digit looks exactly like a real one.
Good enough to save the typing, not good enough to trust blindly. OCR works by recognising shapes, so it confuses similar characters, and a blurry, skewed or low-contrast scan makes that worse. Plan on reading the output against the original, particularly for numbers and names, where an error will not stand out.
This is built for printed text: scans, documents and photos of pages. Handwriting is a different problem, since every writer's letterforms differ, and tools built specifically for handwritten notes handle it better than a general OCR service.
Images as JPG, PNG, BMP, GIF, PCX and TIFF, including multipage TIFF, plus PDFs, including multipage ones. Registered users can also upload a ZIP archive of images to convert together. On the way out a guest chooses Word, Excel or plain text. Registering adds PDF and RTF.
Alternatives to OnlineOCR
Image to Text
Free OCR that pulls the text out of an image, built on the open-source Tesseract engine, with no signup.