Free FineReader alternatives, tested on 10 scanned pages
22 s
searchable PDF and Word file, one OCR pass
ABBYY FineReader is paid, so it did not run. Free tools did two of its jobs instead.
One free tool for each job
| Job | Free tool | Result |
|---|---|---|
| Searchable PDF, Mac | OCRmyPDF 17.13.0 | 1,853 words, 22 s |
| Searchable PDF, Linux | OCRmyPDF Docker image | 1,846 words, 13 s |
| Searchable PDF, Tesseract only | Tesseract 5.5.4 | 1,875 words, 30 s |
| Scan to Word | OCRmyPDF and pandoc | 1,876 words, 22 s |
| Point and click | NAPS2 8.6.0 | 1,879 words, 16 s |
| Most words right | olmOCR-2 7B | 1,954 words, 77 s |
| Plain text | Tesseract 5.5.4 | 1,878 words, 30 s |
| Plain text, no install | Our Actor's code, on our Mac | 1,879 words, 37 s |
All rows ran on Nov 14 except NAPS2 (command line, Nov 13) and olmOCR-2 (our roundup). Times are rough: Mac runs had low priority, Docker did not.
One OCR pass, two files
Install once on a Mac:
brew install tesseract pandoc uv
uv tool install ocrmypdf
Then, in the folder with scan.pdf:
ocrmypdf --jobs 2 --sidecar scan.txt scan.pdf scan.ocr.pdf
perl -0pe 's/-\n(?=[a-z])//g' scan.txt > joined.txt
pandoc joined.txt -o scan.docx
scan.ocr.pdf looks like the scan, with text you can search and copy. It kept all 10 page images as they were: 6.49 MB in and out.
scan.docx has the OCR text in 46 paragraphs. The perl line joins words split at a line end.
On Linux, the OCRmyPDF Docker image makes the PDF:
docker run --rm --cpus 2 -v "$PWD:/data" -w /data \
jbarlow83/ocrmypdf-alpine:v17.13.0 --jobs 2 scan.pdf scan.ocr.pdf
Limits we found
- No layout in Word. The .docx has text only: 0 images and 0 tables. Columns, pictures and fonts stay behind.
- Text read back moves a little. Read the PDF's text back with pdftotext and a few lines come out of order. One "down a" became "downa".
- Linux read 7 fewer words. Its image has Tesseract 5.5.2 and a bigger English model than the Mac's. We did not test which one cost the words.
- NAPS2 shrinks the images. It saved the pages as new JPEGs, 6.5 MB down to 4.2 MB. OCRmyPDF and Tesseract kept them as they were.
Same 10 pages, one scorer
The 10 pages are proofread scans from Wikisource, 1,955 words in all, the same pages as our OCR roundup. The score is true words found, in order, with exact case and punctuation.
How each file was read
- PDF: the text pdftotext gets from it.
- Word file: the text pandoc gets from it.
- Our older pages show 1,855 for OCRmyPDF: the same images in an older PDF copy.
- One Apple Silicon Mac, 2 CPU threads a run, one run at a time.
- Docker ran its own Linux, capped at 2 CPUs. OCRmyPDF ran twice on the Mac and twice in Docker. Each gave the same text on both of its runs.
Scanned pages to text, no install
Run it yourself. Our Apify Actor reads PDF, PNG, JPEG, WEBP and TIFF for $0.006 a page, with nothing to install. Its code found 1,879 of these 1,955 words.
Or send us the scan. Tack, our AI, does the job for $29, up to 50 pages. First page free.
Questions people ask
Is there a free alternative to ABBYY FineReader?
For a searchable PDF, OCRmyPDF found 1,853 of 1,955 words and kept the page images as they were. For a Word file, add pandoc. To click instead of type, NAPS2 has an app.
Is OCRmyPDF as good as FineReader?
We do not know. FineReader is paid and this test spent $0, so FineReader did not run. The free Tesseract tools found 94 to 96% of the words in order. The best vision models found more.
What is a free FineReader alternative for Linux?
OCRmyPDF. Its Docker image found 1,846 of 1,955 words in 13 s and kept every page image byte for byte.
What is a free FineReader alternative for Mac?
OCRmyPDF, with Tesseract and pandoc from Homebrew. The three lines above made both files in 22 s. NAPS2 has a Mac app if you would rather click.