docTR OCR REST API, based on API framework from Mindee
1.3K
Docker image based on the sample API framework by Mindee
Provides OCR capabilities to your applications via REST endpoints.
docTR Github: https://github.com/mindee/doctr
My Github repo for Dockerfile and code: https://github.com/Fybre/doctr-api
I have just added a couple of endpoints for extracting plain text from files, rather than json dictionaries, and made it a bit easier to run with docker compose.
docker-compose.yaml
services:
doctr-api:
image: fybre/doctr-api
container_name: doctr-api
ports:
- 8080:8080
mkdir doctr-api
cd doctr-api
wget https://raw.githubusercontent.com/Fybre/doctr-api/refs/heads/main/docker-compose.yaml
docker compose up -d
For documentation, navigate to http://localhost:8080/docs (or whatever IP and port you deployed to)
The following endpoints are available:
/detection/ - Perform text detection
/ocr/ - Perform OCR, returns json formatted results
/kie/ - Perform KIE
/ocrtext/ - Performs OCR returning text result - result.render()
/ocrtext/single/ - Simple endpoint to perform OCR on a single file, no other parameters needed
/ocrjson/ - Performs OCR, returns simple json with extracted text
/ocrjson/single/ - Simple endpoint to perform OCR on a single file, no other parameters needed. Returns simple json with extracted text
Content type
Image
Digest
sha256:93ea3f4b9…
Size
4.3 GB
Last updated
about 1 year ago
docker pull fybre/doctr-api