OCR capability

The pages of a PDF can be exported to numbered *.png files and using Ollama with Mistral Small 3.2 model works really well. I´m using with my own handwriting notes with minimal manual corrections: ``` ollama run mistral-small3.2:latest "./input.png\nConvierte este documento a formato markdown." ```

Until this is implemented, I recommend using the excellent OCRmyPDF on the PDF, then opening with SumatraPDF. https://ocrmypdf.readthedocs.io/en/latest/index.html