PaddleOCR‑VL Local Web UI (Docker)

Disclaimer
I’m not a programmer, and this entire project was built with the help of AI.
I’d be very happy if others improve it and make it more useful for people.

This repository provides a local web interface for PaddleOCR‑VL.
It runs on localhost, uses Docker for inference, and stores results under data/.

Requirements

• Docker Desktop installed and running
• Image:

  ccr-2vdh3abv-pub.cnc.bj.baidubce.com/paddlepaddle/paddleocr-vl:latest-nvidia-gpu
  

• Python 3.10+

Setup

cd C:\Users\egor0\Downloads\WebforAI\publish
python -m venv .venv
.\.venv\Scripts\Activate.ps1
pip install -r server\requirements.txt

Run

uvicorn server.app:app --host 127.0.0.1 --port 8000

Open in browser: http://127.0.0.1:8000

Outputs

All results are stored in:

data\jobs\<job_id>\output

You can also download ZIP files from the UI.

Notes

• The UI is in English.
• The server always uses PaddleOCR‑VL v1 (no model selection).

Contributing

If you want to improve the UI/UX, performance, or add features (DOCX/PDF export, better previews, queue management), feel free to open issues or PRs.