PyOpenOCR makes it simple to host your own OCR REST API.

The heavy lifting OCR work is handled by Tesseract 4 OCR.

Run the service in your own computer or use Docker to containerize.

• Scalable as a stateless app.
• Platform independence via Docker containers.
• TODO: Kubernetes
• HTTPS out of the box
• Supports 31 languages in addition to English
• Pass arguments to Tesseract such as character whitelist and page segment mode.
• REST API docs

• Install docker
• git clone https://github.com/gabrielpsilva/pyopenocr.git
• cd pyopenocr
• Type ./run_docker.sh (in case you don't have execute right type sudo chmod +x run.sh

• git clone https://github.com/gabrielpsilva/pyopenocr.git
• cd pyopenocr
• Type ./run_native.sh (in case you don't have execute right type sudo chmod +x run.sh

In Tesseract 4.00 The memory usage is light. It is less fast than tesseract 3 and more costly in term of size (an simple aws instance with 1GB of ram is sufficient but with an EBS of 16GB of storage). Result are really better compared to version 3.04. To see a comparative you can have a look to the official page of tesseract

You are now ready to decode images → text via your REST API.

Request with image base64

#convert image to base 64
$ base64 image_test.png > /tmp/image_test.png.txt
$ curl -k -X POST https://localhost:443/ocr-base64 \
  -H "Content-Type: text/plain" \
  -d "@/tmp/image_test.png.txt"

Request with image URL

$curl -k -X POST https://localhost:443/ocr-url \
  -H "Content-Type: text/plain" \
  -d "https://github.com/gabrielpsilva/pyopenocr/blob/master/image_test.png"

Passing parameters to Tesseract

You can pass parameters using in the url request: $curl -k -X POST https://localhost:443/ocr-base64?psm=2&oem=3&lang=eng
-H "Content-Type: text/plain"
-d "https://github.com/gabrielpsilva/pyopenocr/blob/master/image_test.png"

PyOpenOCR is Open Source and available under the Apache 2 License.