Lantern

Privacy

Privacy Policy

Last updated 26 June 2026

Lantern helps communities turn the words they still remember into a course to learn an endangered language. This policy explains, in plain language, what we collect, why, and what you can do about it. You can use the entire app, browse languages, take courses, hear pronunciation, and contribute, without an account. Signing in is optional and only adds saved progress and attribution.

What we collect

  • Account details, only if you sign in: your name, email, and avatar from Google, GitHub, or an email and password you create.
  • Learning progress: which cards you have reviewed and their spaced-repetition schedule, so you can pick up where you left off.
  • Contributions: phrases you add to a language, with the meaning and any source you provide. These become part of that language’s public corpus.
  • Basic technical data needed to serve the app. We do not sell your data or run advertising trackers.

How your data is used and stored

Account and progress data is stored in Firebase and MongoDB Atlas, used only to run the app, save your progress, and credit your contributions. Pronunciation is generated in your own browser with the Web Speech API, so no audio is sent to us. When the engine builds a course, only the language’s cited phrases are sent to our model provider, never your personal information.

Language data and community ownership

Endangered-language data carries real cultural ownership. Our position is that the community owns and governs its corpus, and Lantern is a tool used on that data, never an extraction pipeline. We cite the source of every seed phrase, and formal data-sovereignty controls, letting a community decide who can view, download, and contribute to its language, are an active part of our roadmap. If you represent a community and want your language’s data handled differently, contact us.

Third parties

We use Google and GitHub for optional sign-in, Firebase and MongoDB for storage, and a hosted language model for course induction. Each receives only the data needed for its function, and never your personal data for model inference.

Your choices

You can use Lantern fully without an account. If you have one, you can ask us to delete your account and the personal data tied to it. Public contributions may remain in a language’s corpus to preserve the community record, but we will detach them from your identity on request.

Contact

Questions about privacy or a community’s data can be sent to the maintainers via the GitHub repository.