Extract text from any
image or document free
Drop a photo, screenshot, or PDF — get copyable text in seconds. Free, no signup needed.
Drop image or PDF here
or click to browse · paste from clipboard ( ⌘V)
No signup required · 20 pages/day free with account · 100+ languages
Upload
Drop an image, PDF, or paste from clipboard instantly.
Extract
Get editable text in seconds — supporting 100+ languages.
Use
Copy, download, or share a link to the result immediately.
Works with anything you throw at it.
Photos, scans, screenshots, handwriting — if it has text, we can extract it.
Photos of documents
Take a photo of any paper, receipt, or page with your phone. Get the text instantly — no scanner needed.
Handwritten notes
Convert handwriting to typed text you can edit, search, and share. Works with cursive and print.
Scanned PDFs
Make any scanned PDF searchable and copyable. Process multi-page documents in one go.
Screenshots & screen captures
Pull text from any screenshot instantly. Perfect for error messages, code snippets, and web content.
Any document. Editable text.
Receipts, contracts, textbooks, notes — drop it in, get the text out.

Receipts & invoices
Expense tracking made easy

Contracts & letters
Turn paper into editable text

Books & articles
Digitize printed pages fast
A tool for every task
Jump straight to the right tool for your use case, or browse all tools.
100+ languages supported
OCR Performance Benchmarks
Real accuracy numbers across industry-standard datasets. See how our models stack up.
PP-OCRv5
2M params| Dataset / Task | PP-OCRv5 | Google Vision | AWS Textract |
|---|---|---|---|
| ICDAR 2015 · Printed English | 82.5% F1 | 84.2% | 83.1% |
| MSRA-TD500 · Multilingual Scene | 87.3% F1 | 85.1% | — |
| Chinese Document · Dense CJK | 95.2% | 93.8% | 89.5% |
| English Document · Standard | 97.1% | 97.5% | 96.8% |
| Handwritten · English | 78.4% | 82.1% | 76.3% |
| Handwritten · Chinese | 81.2% | 79.5% | — |
| Receipts & Invoices · Structured | 92.8% | 94.1% | 95.2% |
Key advantage: CJK recognition outperforms Google Vision. 2M parameter model enables fast local inference.
PaddleOCR-VL
| Task | PaddleOCR-VL | Google Doc AI | AWS Textract |
|---|---|---|---|
| Simple Tables | 91.5% | 93.2% | 94.1% |
| Complex Tables | 84.3% | 86.7% | 85.9% |
| Mixed Layouts | 89.1% | 91.4% | 88.5% |
| Handwritten + Printed Mixed | 83.7% | 85.2% | 80.1% |
Google sign-in and your data
When you choose Google sign-in, ScanRead.ai requests only openid, email and profile permissions. We receive your Google account identifier, email address and available basic profile information, such as your name. These permissions do not grant access to your Gmail messages, Google Drive files, contacts or calendar.
We use this information to create or link your ScanRead.ai account, sign you in and identify your account. Authentication is handled through our sign-in service at auth.linnk.ai. We do not sell Google user data or use it for advertising or AI model training.
Privacy PolicyFrequently Asked Questions
Simple, transparent pricing
Free forever. Upgrade when you need more.
View pricing plans