A modern, BYOK and FOSS android app to summarize videos(YouTube, BiliBili), articles, images and documents with AI/LLM.
MAD: pure Kotlin + Jetpack Compose + M3 Expressive
Minimal Android version: 13
There are 2 releases available:
-
standalone which includes ML model to recognize text from image, it got larger package size.
-
gms which allows user to download the model from Google Play Services, but GMS required.
The Play Store release bundled with gms version. This package is signed by Google managed key for simplicity, which means it's not compatible to update in Google Play Store if you installed a github package previously.
- Summarize multiple media types
| media | supported types |
|---|---|
| Video (Link) | YouTube, BiliBili |
| Document | MS Word, PDF |
| Image | Jpg, Png, Webp (Latin only for now) |
| Text | Article link, Plain text |
-
Multiple LLM providers supported
- OpenAI Chatgpt
- Google Gemini
- Anthropic Claude
- Alibaba Qwen
- DeepSeek
- Mistral
- OpenRouter
- Ollama
-
Material 3 Expressive UI: Engaging and easier to use, light/dark theme and dynamic color theme
-
Instant summarize via share sheet or text selection toolbar: Convenient to trigger summarization, show result in overlayer
-
Configurable LLM settings
-
Text to speech for the summaries with multilingual and speed adjustment support
-
History search
-
The original idea from talosross
-
Koog for kotlin-based LLM interactions