yasahiro10/scanner_v3 repository overview

🖨️ Scanner OCR Docker Image

This Docker image packages a complete document scanning and OCR workflow into a single container, allowing you to scan physical documents, extract text, analyze it with a local LLM, normalize content, and save results to multiple databases. Perfect for invoices, purchase orders, and quotes.

Features Scan documents via SANE-compatible scanners (Brother brscan5 supported) OCR extraction using external services Text analysis with a local LLM (qwen2.5:14b via Ollama) Article normalization using Elasticsearch Structured data persistence in MySQL and CouchDB Visualization through a Kibana dashboard Simple Tkinter GUI for scanning and managing documents Quick Start

Run directly from Docker Hub:

docker pull yasahiro10/scanner_v3 docker run --rm -it
-e DISPLAY=$DISPLAY
-v /tmp/.X11-unix:/tmp/.X11-unix
--privileged
yasahiro10/scanner_v3

Build locally:

git clone cd scanner_v3 docker build -t scanner_v3 . docker run --rm -it
-e DISPLAY=$DISPLAY
-v /tmp/.X11-unix:/tmp/.X11-unix
--privileged
scanner_v3

⚠️ --privileged is required for USB scanner access.

Supported Paper & Image Formats Paper: A4, A5, A6, B5, Auto-detect Image: PNG, JPEG, TIFF, BMP Configuration

Override defaults by creating config.py in /app for CouchDB, Kibana, or other endpoints.