Active development: OCR-derived information may be inaccurate.
Fast view of cheapest listings

Help out

You can help this project by running the OCR worker on your machine. It connects to RabbitMQ, processes OCR tasks, and sends results back.

Step 1: Install Docker

  1. Install Docker Desktop for Windows / Mac, or Docker Engine on Linux.
  2. Make sure Docker is running (you can open Docker Desktop and confirm “Running”).

Step 2: Create a folder with a compose file

Create a new folder (example: quinfall-worker) and add a file named docker-compose.yml with this:

services:
  ocr_consumer:
    image: akjroller1995/quinfall-ocr-consumer:latest
    container_name: quinfall_ocr_consumer
    restart: unless-stopped
    environment:
      RABBITMQ_HOST: rabbitmq.therollermethod.com
      RABBITMQ_PORT: 5672
      RABBITMQ_USER: ocr_client
      RABBITMQ_PASS: guest1234
      RABBITMQ_INPUT_QUEUE: ocr_tasks
      RABBITMQ_OUTPUT_QUEUE: ocr_results

  watchtower:
    image: containrrr/watchtower
    container_name: quinfall_watchtower
    restart: unless-stopped
    volumes:
      - /var/run/docker.sock:/var/run/docker.sock
    command: --cleanup --interval 300 quinfall_ocr_consumer

Notes: Watchtower checks every 5 minutes (300s) and automatically updates the worker if a new image is published.

Step 3: Run it

In that folder, run:

docker compose up -d

To check logs:

docker logs -f quinfall_ocr_consumer

To stop:

docker compose down

Safety / permissions

  • This worker only needs access to the task queues (no admin access).
  • If you ever rotate credentials, update them in the compose file.
  • Watchtower needs Docker socket access to update containers.