Text Extractor
Extract text from images, PDFs, Office files, EPUB, and subtitles — entirely in your browser.
About the text extractor
Free online text extractor and OCR. Pull text from images, scanned PDFs, DOCX, PPTX, XLSX, EPUB, SRT, and more. Runs in your browser — no sign-up, files stay on your device.
What you can do
Every capability in this tool, in plain language.
Read words in pictures
Optical character recognition (OCR) for PNG, JPEG, WEBP, BMP, GIF, and PBM — including screenshots you paste from the clipboard (Ctrl/Cmd+V).
Pull text out of PDFs
PDF to text from the embedded text layer, with an optional OCR pass for scanned pages that have no selectable text. Turn “OCR pages with no text layer” off for faster digital PDFs.
Office files and e-books
Word (.docx), PowerPoint (.pptx), Excel (.xlsx), OpenDocument (.odt, .odp, .ods), plus EPUB e-books — extract the words without uploading the file.
Captions and everyday files
Subtitle transcripts from .srt and .vtt (timestamps stripped), plus HTML, RTF, Markdown, CSV, TSV, JSON, XML, YAML, logs, and common source files (JS, TS, Python, and more).
Pick a language for OCR
OCR languages include English, Spanish, French, German, Italian, Portuguese, Dutch, Polish, Turkish, Russian, Arabic, Hindi, Simplified and Traditional Chinese, Japanese, Korean, and Vietnamese. The language model downloads once and is cached in your browser.
Files stay on your device
Runs entirely in your browser — files are never uploaded. For images and scanned PDFs, Tesseract.js fetches an OCR language model from a public CDN the first time you use a language; your documents themselves never leave the device.
Several files in a row
Drop many files at once, or keep adding them. Each file is queued, extracted in turn, and listed so you can jump between results.
Fix mistakes, then copy
Select a file to review the extracted text. Edit OCR mistakes, copy the text, download a .txt per file, or save every result as a ZIP. Character and word counts stay in the status bar.
Big files are OK
Each file can be up to 80 MB. Drag and drop onto the workspace, click Open, or paste a screenshot.
How to use it
- 1
Drop files onto the workspace, click Open, or paste a screenshot (Ctrl/Cmd+V).
- 2
Pick an OCR language if you are extracting from an image or a scanned PDF.
- 3
Leave “OCR pages with no text layer” on if you want scanned PDFs read as images; turn it off for faster digital PDFs.
- 4
Select a file in the list to review the extracted text. Edit it if OCR made a mistake.
- 5
Copy the text, download a .txt, or download every finished file as a ZIP.
Frequently asked questions
Are my files uploaded to a server?
No. Each file is read in your browser. Nothing is sent to us. For images and scanned PDFs, Tesseract.js fetches an OCR language model from a public CDN the first time you use a language; that model is cached in your browser afterwards. Your documents themselves never leave the device.
Which file types can I extract text from?
Images (PNG, JPEG, WEBP, BMP, GIF), PDFs, Word/PowerPoint/Excel Office files (.docx, .pptx, .xlsx), OpenDocument (.odt, .odp, .ods), EPUB, subtitles (.srt, .vtt), HTML, RTF, and everyday text formats such as TXT, Markdown, CSV, JSON, XML, and YAML.
Does it work on scanned PDFs and screenshots?
Yes. Screenshots and photos go through OCR. For PDFs, digital documents use the text layer; if a page has no text and “OCR pages with no text layer” is on, that page is rasterized and read with OCR.
Can it transcribe audio or video?
No. Speech-to-text would require a large on-device model or a remote API. If you already have captions, drop an .srt or .vtt file and the tool will strip timestamps so you get a clean transcript.
Does the text extractor work on mobile?
Yes. You can open files from your phone or tablet. OCR of large scans is slower on small devices because all the work happens locally.