This small Docker container allows to run Tesseract OCR on PDF documents via HTTP POST. To do so, a PDF document is submitted as "userfile" form data. The container returns a PDF document with 300 DPI resolution and all recognized text included.
docker build -t ocr .
docker run -p8080:80 ocr
curl -F "userfile=@/tmp/test.pdf" -H "Expect:" -o output.pdf localhost:8080/ocr.php
Currently, Tesseract settings like language, etc. are hard-coded. These should be made configurable via the REST interface.