bi4key

Ok, thx for your work! Here some inspiration for future text recognition and summary: https://github.com/docling-project/docling https://huggingface.co/ibm-granite/granite-docling-258M https://huggingface.co/unsloth/gemma-3-270m-it-qat-GGUF

And here is nice mega thread: https://www.reddit.com/r/LocalLLaMA/s/nBLTtkVPvK https://huggingface.co/blog/ocr-open-models

In theory in new update you can set "auto detection" with new "bigger model", and app auto find language