Winrecall 0.25 - Preview

Initial Release

This first release has two image analysis modes:

• OCR Only uses Tesseract OCR for fast and simple text extraction, perfect if you need quick results.
• AI + OCR combines traditional text recognition with a deep learning model (the BLIP model from Hugging Face) to generate detailed descriptions based on image content. Processing time varies: AI + OCR takes approximately 10–40 seconds per image, depending on hardware. For example, on an i7-9700 CPU with 32 GB RAM, processing takes around 10 seconds. If speed is your priority, choose OCR Only. If you prefer AI-powered insights, AI + OCR is the better option.
• In this release, snapshots are now encrypted. You'll need to set a session key to log in and decrypt the images. This key is valid only as long as the database exists. If the database is reset, a new key will need to be created.

For the AI model, make sure to run the setup environment as outlined in the README. OCR comes pre-packaged with the app and includes both English and German training data.

My first thought for the app name was 'PeepCall' as a little jab at Microsoft’s Recall, by the way 😄

Keep in mind that it’s still a work in progress, so some bugs or glitches might pop up. Feel free to test it out, and don’t hesitate to share your feedback!