OCR Text Extractor: text from images and scanned PDFs
Pricing
from $20.00 / 1,000 page processeds
OCR Text Extractor: text from images and scanned PDFs
Reads the text in images (PNG, JPEG, WebP, TIFF) and scanned PDF files you link to, page by page, with a confidence score for every line. OCR runs inside the actor; no outside service is called.
Pricing
from $20.00 / 1,000 page processeds
Rating
0.0
(0)
Developer
Ed Rayyis
Maintained by CommunityActor stats
0
Bookmarked
2
Total users
1
Monthly active users
8 days ago
Last modified
Categories
Security
Limited permissionsShare
Give it links to images or PDF files and it returns the text on every page, read with optical character recognition (OCR). Each page comes with a mean confidence score, and you can ask for every line with its own confidence and position.
- Images: PNG, JPEG, WebP, BMP, GIF, and multi page TIFF (each frame is a page)
- PDF files: each page is rendered and read. Pages that already carry a text layer can return that text directly, without OCR
- Tested on English, German, French, Italian, Spanish and Portuguese text, and on a typewritten scan from 1946
OCR runs inside the actor with RapidOCR and the PP-OCRv6 models. No outside OCR service is called and no API key is needed.
Input
- Image or PDF URLs (required), up to 50 MB per file
- Use the PDF text layer when present (default on)
- Maximum pages per file (default 50)
- PDF render resolution in DPI (default 200)
- Minimum line confidence in percent (default 50): lines below it are dropped
- Include line details (default off): every line with its confidence and bounding box
Output
Dataset: one item per file with url, ok, fileType (pdf or image), pageCount, pagesProcessed, text (all pages), and pages. Each page has page, method (ocr or text-layer), text, meanConfidence, and with line details on, lines. A file that cannot be downloaded or read gets ok: false and an error. Export as JSON, CSV or Excel, or read it over the API.
Pricing
Pay per event: one event per page processed, whether read by OCR or from the text layer. Set the page limit to control cost. The run stops when it reaches the maximum charge you set, and the file it stopped in is marked with a note.
Scope and conduct
Only the URLs you supply are fetched. Files are processed in memory and nothing is kept beyond the run's own dataset. Handwriting was not tested.
Apify gives a run one CPU core for every 4 GB of memory, so OCR speed follows the memory you choose. A full page of small print (a US tax form) took about 20 seconds with 4 GB.
Support
Questions and problems: the Issues tab on this page. Replies are written by an AI assistant on behalf of the developer and reviewed by the developer.
