Image to text for Bulgarian and English: free online OCR
Choose a photo, a scanned document or a PDF (its first page is read) and get the text, ready to copy, edit and download. It reads printed text in Bulgarian (Cyrillic) and English. The image is never uploaded: it all runs in your browser.
Opens without an account 3 times a day. A free account removes the limit.
- Image to textto copy, edit and download
- Bulgarian and Englishprinted text in Cyrillic and Latin
- Nothing is uploadedrecognition runs in your browser
- Searchable PDFinvisible text over the page
What image to text (OCR) is
Image to text (OCR, optical character recognition) turns the letters in a picture into text you can copy, edit and search. The EduFacturing tool does it online and free for a photo, a scanned document or a PDF, in Bulgarian and English.
Before recognition you see whether the image is fit to read and exactly what the program will read.
The image is never uploaded: it is processed in your browser by Tesseract, an open-source OCR engine.
Image to text: what the tool does
Extracts text in Bulgarian and English
You choose Bulgarian, English or both together. The language data downloads on the first run (on screen 1.6 MB for Bulgarian, 3.9 MB for English and 5.5 MB for both), and the browser keeps it after that.
Reads a scanned PDF
The first page of a PDF is rendered as a picture at 300 dpi and recognised like a photo.
Prepares the image for better OCR
It shows size, contrast and brightness and warns when the height is under 800 px. It turns the image black and white with an automatic threshold (Otsu's method) or a local threshold for shadows, and can invert black and white.
Gives text, word confidence and a searchable PDF
The result is a text box that you edit, copy and download as .txt. You see how many words were read and what share of them with confidence, and the uncertain ones are listed. The "Download searchable PDF" button gives a PDF in which Ctrl+F finds the words. For Bulgarian, Latin letters that look like Cyrillic ones are replaced with Cyrillic.
When you need text from an image or a scanned document
You need image to text when the text exists only as a picture and you want to edit, search or quote it.
Scanned document to text or to a searchable PDF
A report, a contract or a letter in Bulgarian, scanned as a picture or a PDF, becomes editable text or a searchable PDF.
Text from a photo of a textbook or notes
You photograph a textbook page or class notes and get the text instead of retyping it.
Copy text from a slide, a sign or a sheet
Text from a photo of a slide, a sign or a sheet is copied and downloaded as .txt.
How the text recognition works
Recognition takes five steps: assessing the image, preparing it, reading the letters, correcting them and scoring the words.
Assess the image
Size, contrast (how much the light and dark areas differ) and mean brightness are worked out from the pixels.
Prepare
The image is turned grey. The automatic threshold (Otsu's) picks one grey level for the whole page, and the local one compares each pixel with the mean around it, which helps with shadows.
Recognise the text
Tesseract (through tesseract.js 7, with an LSTM neural network) reads the prepared picture and returns the text.
Correct the letters
For Bulgarian and for both languages, a Latin letter is replaced by its Cyrillic look-alike inside a word in which Cyrillic letters are at least as many as Latin ones. An English word in the text is left alone.
Confidence for every word
Every word comes with a confidence from 0 to 100. The screen shows what share of the words have at least 70% and warns when it is under 80%. Words with confidence under 30% are left out of the searchable PDF.
| Applies to | Formula | Meaning |
|---|---|---|
| Otsu threshold | t = argmax w₁·w₂·(μ₁ − μ₂)² | the grey level that best separates text from background |
| Local threshold | t(x, y) = mean(31 × 31 px) − 8 | a pixel is white above t(x, y); the window has a radius of 15 px |
| Warnings | height < 800 px; contrast < 35; brightness < 55 or > 215 | each condition gives its own message before recognition |
Free Bulgarian OCR online: what is different here
You see what the program will read
The “Show what the recognition will see” button draws the prepared picture, so a shadow or a black blotch shows before you wait.
It warns about a poor image before you wait
Size, contrast and brightness are assessed as soon as you choose the file.
It tells you which words not to trust
Words with confidence under 70% are listed separately, up to 60 at a time, so you check those first.
It reads printed text in Bulgarian and English. It does not read other languages or handwriting. Columns and tables are lost: the text comes out as one stream.
A tilted photo is not straightened and is recognised worse. Only the first page of a PDF is read. The result is text (.txt) or a searchable PDF, not a Word file. In that PDF the page is the prepared grey or black-and-white picture, not the colour original.
Questions about image to text
How to extract text from an image?
Choose the image, pick Bulgarian, English or both, and start the recognition. The text appears in a box, from which you copy it or download it as .txt.
Is there a free Bulgarian OCR online?
Yes. This tool reads Bulgarian (Cyrillic) and English. Without an account it opens 3 times a day, and a free account removes the limit.
How do I copy text from a picture into Word?
Recognise the picture, copy the text from the result box and paste it into Word. The tool downloads the text as .txt, not as a Word file.
How do I extract text from a scanned PDF?
Choose the PDF instead of a picture. Its first page is rendered as a picture at 300 dpi and recognised. Then "Download searchable PDF" gives a PDF in which Ctrl+F finds the words.
Is my image uploaded anywhere?
No. Recognition runs in your browser and the image is not uploaded.
Why does the recognised text have mistakes?
Most often because of a small, dark or tilted photo. Under 800 px of height the tool warns you, and with a shadow the local threshold helps. Words with confidence under 70% are listed below the result (up to 60): check those first.
Image to text is free. Pro opens the rest
Pro costs 15 EUR a month and opens 45 Pro tools.
In the Design group:
Related tools
Document ScannerScan a document from a photo, free: click the four corners of the page and download a straightened A4 or Letter page as a PNG or a PDF.
PDF tools: merge, split, shrinkMerge PDF without uploading: join files, split pages, shrink, number, add a watermark, lock and sign. Your file stays in the browser. Pro, 15 EUR a month.
DOCX to Markdown, HTML and PDFDOCX to Markdown online: a Word or LibreOffice document becomes Markdown, HTML, plain text or PDF. Free, 3 times a day without an account, no limit with one.
Image converter and resizerFree online image converter: JPG to PNG, WebP to JPG, HEIC to JPG. Reduce image size to a target in KB, crop and strip GPS. Runs in your browser, no upload.
Extract the text from your image
Opens without an account 3 times a day. A free account removes the limit.