Paddle Drive OCR POC

A local upload/download/search proof of concept with a Drive-style UI and PaddleOCR-backed text indexing.

Run

python3 -m pip install -r requirements.txt
python3 app/server.py

Then open http://127.0.0.1:8765.

Uploaded files are stored in uploads/. Metadata and OCR text are stored in data/paddle_drive.sqlite3.

Notes

• Image files (png, jpg, jpeg, webp, bmp, tiff) are queued for PaddleOCR.
• Word (docx) and Excel (xlsx) files are indexed by extracting their embedded text directly.
• Other files are still stored and downloadable, but marked as download-only.
• PaddleOCR may download model files on first use.