OCR - Extract Text Image PDF Free

OCR

Pull text out of images and scanned PDFs. 100+ languages, line and word boxes, layout-preserving export. Output text, hOCR, TSV, DOCX, or a fully searchable PDF. Powered by Tesseract.js, runs entirely in your browser.

100% in your browser No upload 100+ languages Searchable PDF

Input drag images or PDFs

Drop images, scans, or PDFs here

or click to browse. PNG, JPG, WEBP, BMP, TIFF, PDF

Settings

Languages (Ctrl-click for multi)

Page segmentation Auto (default) Auto with orientation detect Single column variable size Single uniform block Single text line Single word Sparse text Sparse text with orientation Raw line PDF render scale 2x (faster) 3x (recommended) 4x (highest, slow) 5x (very slow)

Model quality Best accuracy (~15 MB / lang) Standard (~5 MB / lang) Fast (~2 MB / lang) Engine mode LSTM only (recommended) LSTM + legacy Legacy only

Image preprocessing None (best for clean docs) Grayscale only Binarize Otsu (faded scans) Invert + binarize (white on dark) Upscale 2x + sharpen Auto-deskew (rotation correction) On (PSM 1 with OSD) Off

Idle

Output awaiting OCR

Run OCR to detect tables.

Blocks Lines Words

Run OCR to view annotated image.

Page 0 / 0

Languages

First-time use of a language downloads its trained data (~1-15 MB each) from jsDelivr. Cached after that. Multi-select to OCR mixed-script documents in one pass.