Repository navigation
First public release.
Dear is a second brain that lives on your Android phone: share links, photos, videos, voice notes or text into it, and Gemma 4 catalogs them on-device (title, summary, tags, text in images, transcripts). Then ask your collection questions in plain language, with answers that cite their entries.
What's in 0.1.0
- Share target for links, photos, videos, audio and text; notes with the + button
- Multimodal cataloging with Gemma 4 E2B / E4B through llama.cpp + libmtmd, in the background
- Public Instagram posts: carousel photos pulled in and cataloged with the link
- Collection by day with keyword search and type filters
- Ask: chat with your collection, with citations
- Export / import of the whole collection as a .zip
- English and Italian interface
Install
- arm64 Android 8.0+; 6 GB RAM or more for E2B (recommended), 8 GB for E4B
- Install
dear-0.1.0.apk, then download a model from Settings (about 3.4 GB for E2B, once, from Hugging Face)
Privacy: no account, no telemetry, no cloud. The app goes online only for the model download and to read the metadata of the links you save.
APK SHA-256: 9338255d6709f588dc53068d0eaf668fa7896ea417be292a7c3bc1650e6ab12a