Scan → Ollama → PDF

A desktop app that sends scanned handwritten documents directly to a local Ollama vision model (default: gemma4) and saves the transcription as a plain-text PDF.

Requirements

• Python 3.10+
• Ollama running locally with a vision-capable model pulled (e.g. ollama pull gemma4)

Installation

pip install requests pillow fpdf2 pypdfium2

Usage

python handwriting_app.py

1. Click Add Images… to queue one or more scanned image files (PNG, JPG, BMP, TIFF), or Add PDF… to queue a scanned PDF.
2. Select the Ollama model from the dropdown (auto-populated from your local instance).
3. Click Transcribe & Save PDF…, choose an output path, and wait while each page is processed.
4. The transcription is shown in the preview pane and saved as a PDF.