OCR — Image to Text Extractor
Pull text out of screenshots, scans and photos — self-hosted engine, image never uploaded
Recognition runs entirely in your browser via a self-hosted Tesseract engine (downloaded once, then cached). Your image never leaves the device.
Examples
Quote from a scan
Scanned paragraph photo, 300 px tall≈ verbatim text · confidence 88%Contrast stretching keeps thermal-printed and shaded text recognizable.
About this tool
Tesseract compiled to WebAssembly, running fully in your browser with a self-hosted engine and English model (~13 MB, downloaded once and cached — offline after that). Before recognition, images are auto-enhanced: grayscale plus contrast stretching, which markedly lifts accuracy on photos and receipts.
Results arrive with a confidence score, in an editable pane with copy and .txt export. Honest limits: handwriting and stylized display fonts stay unreliable, and sharp straight-on input remains the recipe.
How to use
Pick an image
PNG, JPG, WebP or BMP up to 20 MB.
Extract
Engine loads once (cached); live progress streams while it recognizes.
Use the text
Copy or download; check confidence and compare with the preview.
Use cases
Screenshot salvage
Code and quotes out of talk slides and video calls.
Receipt digitization
Line items into text for expense notes.
Common mistakes
Mistake:Feeding blurry or angled photos
Fix:OCR rewards sharp, straight-on, high-contrast input; retake instead of hoping.
Mistake:Trusting low-confidence output verbatim
Fix:Below ~80 confidence expect character swaps — proofread against the preview.
Frequently asked questions
Related guides
Text to Speech in the Browser: Voices, Chunking and the No-Download Truth
How speechSynthesis works on-device, why long texts get split at sentence boundaries, and the honest reason TTS can't export audio files.
PDF to Images: DPI, Page Ranges and a Self-Hosted Rendering Engine
How browser-side PDF rasterization works at 72–300 DPI, why the pdf.js worker is self-hosted, and the compact range syntax worth knowing.
Your Photos Know Where You Live: Stripping EXIF and GPS Metadata
What EXIF, GPS and PNG text chunks reveal about you, and how byte-level metadata stripping produces a shareable clean copy.
References & standards
Related tools
LLM Benchmarks
Compare LLM coding, reasoning, speed and cost benchmarks side-by-side and get a task-specific model recommendation
Video to GIF Maker
Cut a clip from any video and convert it to a high-quality GIF — palette-optimized colors, size and FPS control, fully local.
Currency Converter
Convert between 30 currencies using an offline rate snapshot with manual rate override — no API calls, works without a network.
Text to Speech (TTS)
Read any text aloud with your browser's built-in voices — control voice, rate, pitch and volume, all processed on your device.
Voice Dictation
Dictate with your voice and watch the text appear live — 12 languages, browser speech engine, transcript stays on your device.
Ambient Noise Generator
Blend white, pink and brown noise live in your browser — focus masking, tinnitus relief and sleep, with a built-in sleep timer.