Third-Party Notices
Last updated: 2026-08-26
Eduvia (the iOS app and its backend service) includes the open-source components listed below. We are grateful to their authors and maintainers.
Exact version numbers are pinned in the project lockfiles (Package.resolved for iOS, package-lock.json for the server) and are read from there at each release.
Voice synthesis (server side)
| Component | License | Source |
|---|---|---|
| Piper (piper1-gpl) | GPL-3.0 | https://github.com/OHF-Voice/piper1-gpl |
| eSpeak NG (used by Piper for phonemization) | GPL-3.0 | https://github.com/espeak-ng/espeak-ng |
| ONNX Runtime | MIT | https://github.com/microsoft/onnxruntime |
| Voice model de_DE-thorsten-medium | Training dataset released under CC0 by the Thorsten-Voice project; distributed via rhasspy/piper-voices; per its model card, finetuned from the U.S. English lessac voice | https://github.com/rhasspy/piper-voices and https://www.thorsten-voice.de |
Server
| Component | License | Source |
|---|---|---|
| Fastify | MIT | https://github.com/fastify/fastify |
| @fastify/cors | MIT | https://github.com/fastify/fastify-cors |
| ws | MIT | https://github.com/websockets/ws |
| @google/genai | Apache-2.0 | https://github.com/googleapis/js-genai |
| firebase-admin | Apache-2.0 | https://github.com/firebase/firebase-admin-node |
| dotenv | BSD-2-Clause | https://github.com/motdotla/dotenv |
iOS app
| Component | License | Source |
|---|---|---|
| FluidAudio | Apache-2.0 | https://github.com/FluidInference/FluidAudio |
| Silero VAD (voice activity detection model, bundled with the App) | MIT | https://github.com/snakers4/silero-vad |
| Firebase Apple SDK (FirebaseCore, FirebaseAuth, FirebaseCrashlytics, FirebaseInstallations, FirebaseSessions) | Apache-2.0 | https://github.com/firebase/firebase-ios-sdk |
| FirebaseAnalytics / GoogleAppMeasurement (closed-source Google binaries used by Firebase) | Google APIs Terms of Service | https://developers.google.com/terms |
| RevenueCat (purchases-ios) | MIT | https://github.com/RevenueCat/purchases-ios |
| GoogleSignIn-iOS | Apache-2.0 | https://github.com/google/GoogleSignIn-iOS |
| AppAuth-iOS | Apache-2.0 | https://github.com/openid/AppAuth-iOS |
| GTMAppAuth | Apache-2.0 | https://github.com/google/GTMAppAuth |
| GTMSessionFetcher | Apache-2.0 | https://github.com/google/gtm-session-fetcher |
| GoogleUtilities | Apache-2.0 | https://github.com/google/GoogleUtilities |
| GoogleDataTransport | Apache-2.0 | https://github.com/google/GoogleDataTransport |
| AppCheckCore | Apache-2.0 | https://github.com/google/app-check |
| Promises (FBLPromises) | Apache-2.0 | https://github.com/google/promises |
| nanopb | zlib | https://github.com/nanopb/nanopb |
Notes
- The Silero VAD model ships inside the App bundle; nothing is downloaded from third parties at runtime.
- GPL-3.0 licensed components (Piper, eSpeak NG) run exclusively on the server side and are not distributed to end users as part of the App. GPL-3.0 source obligations attach to the distribution of binaries.
- The de_DE-thorsten-medium voice model card states that the training dataset is CC0 and that the model was finetuned from the U.S. English lessac voice.
- Full license texts are available at the linked repositories.