Open-source licenses
Last updated
Echo runs entirely on your phone thanks to open models and open-source software. We are grateful to the people who build them. Follow each link for the full licence text.
AI models in the app
| Component | Used for | Licence |
|---|---|---|
| Whisper (Large V3 Turbo, Medium, Small, Base) | Speech recognition | MIT |
| Qwen3 1.7B and 0.6B | Notes AI | Apache 2.0 |
| Gemma 4 E2B | Multilingual Notes AI (optional) | Apache 2.0 |
| Silero VAD | Voice activity detection | MIT |
| pyannote segmentation 3.0 | Speaker detection | MIT |
| 3D-Speaker ERes2Net | Speaker detection | Apache 2.0 |
Libraries in the app
| Component | Used for | Licence |
|---|---|---|
| whisper.cpp / ggml | On-device Whisper engine | MIT |
| sherpa-onnx | Speaker detection engine | Apache 2.0 |
| LiteRT-LM | On-device Notes AI engine | Apache 2.0 |
| Android Jetpack (AndroidX, Compose, Media3, WorkManager, Glance, Navigation) | App framework | Apache 2.0 |
| SQLDelight | Database | Apache 2.0 |
| kotlinx.serialization | Data format | Apache 2.0 |
| RevenueCat Purchases SDK | Pro purchases | MIT |
