Any URL OCR
Under maintenancePricing
from $50.00 / 1,000 ocr per pages
Any URL OCR
Under maintenanceExtract text from any image or PDF by URL using Tesseract OCR. Multi-language, PDF page-by-page, no paid OCR API.
Pricing
from $50.00 / 1,000 ocr per pages
Rating
0.0
(0)
Developer
Caleno
Maintained by CommunityActor stats
0
Bookmarked
2
Total users
1
Monthly active users
21 hours ago
Last modified
Categories
Share
URL OCR 🔤
Extract text from any image or PDF by URL — automatic language detection, PDF support, no paid OCR API.
Point it at an image or PDF URL and get the text back. It uses a modern deep-learning OCR engine, so it reads the hard cases too: screenshots, UI and menus, photos, low-contrast scans and light-text-on-dark backgrounds — not just clean black-on-white documents.
✨ What it does
- 🖼️ Images & 📄 PDFs — PNG, JPG, WebP, TIFF, and multi-page PDFs (processed page by page).
- 🌍 Automatic language detection — no need to pick a language. English, Hungarian, German, French, Spanish, Italian, Portuguese, Dutch, Polish and more.
- 🌓 Dark-background aware — automatically inverts light-text-on-dark images (navbars, hero sections, dark UI) that plain OCR reads as garbage.
- 🎯 Clean output — plain text plus word/character counts and the detected language; PDFs also return per-page text.
- 💸 No external OCR API — everything runs inside the actor, keeping it cheap.
🎯 Use cases
- Digitize scanned documents, invoices and receipts
- Pull text out of screenshots, app UIs and social posts
- Make PDFs searchable (page-by-page text)
- Feed extracted text into search, AI, translation or data pipelines
- Read text from product photos and signage
📥 Input
| Field | Description |
|---|---|
| Image / PDF URLs | One or more image or PDF URLs to read. |
| PDF render quality (DPI) | Resolution for rasterizing PDF pages (default 200). |
| Proxy | Optional proxy for downloading the files. |
Example input
📤 Output
One dataset item per URL:
For a PDF you also get pageCount and a pages array with the text of each page.
⚙️ How it works
The file is downloaded, images are preprocessed (upscaled and cleaned up so small or low-contrast text is readable), and the text is recognized with a deep-learning OCR model. The language is detected from the result automatically. PDFs are rasterized page by page and each page is read separately.
💳 Pricing
Pay per page: you are charged per page processed — an image counts as one page, a PDF is charged per page. Failed URLs are free.
❓ Notes
- Best results come from real text. Stylized logos and heavy artwork are graphics, not text, so they may not read cleanly.
- Very low-resolution or blurry sources reduce accuracy — a higher-quality image gives better text.
Questions or a feature request? Open an issue on the actor page — the actor is actively maintained. 🔄
![Google Ads Bulk Scraper [NO COOKIES] ✅ avatar](https://images.apifyusercontent.com/5oFn7z_yEEqtiW_e5j2B9c85BQyFGAPaFmylecDsK6M/rs:fill:76:76/cb:1/aHR0cHM6Ly9hcGlmeS1pbWFnZS11cGxvYWRzLXByb2QuczMudXMtZWFzdC0xLmFtYXpvbmF3cy5jb20vc0NDU2R3bm1LTUoxUTJzVG4tYWN0b3ItcGtKbVNWQkk4M3ZGeXkycjUtVmp6Tm44SzRHVi1wbmctY2xpcGFydC1nb29nbGUtbG9nby1nb29nbGUtbG9nby1nLXN1aXRlLWNocm9tZS10ZXh0LWxvZ28tdGh1.webp){kind=avatar}
