Last updated · 2026-05-25

Privacy Policy

Last updated: 2026-05-25

Thank you for helping us build a Kashmiri speech-to-text (STT) dataset. This policy covers both the mobile app and the browser-based web recorder at kashmiristt (collectively, "the app").

What we collect

When you use the app, we may store the following:

  • Profile information you provide during onboarding: name, age, gender, district, native language, education.
  • Account identifier: a Google account ID (sub) if you sign in with Google, an Apple identifier if you sign in with Apple, or an employee code issued by our team.
  • Voice recordings you create by reading the prompts shown in the app.
  • Recording metadata: prompt ID, duration, sample rate, device type, timestamp.

Employees who sign in with an employee code do not re-enter profile information — their profile is associated with the code by our team in advance.

How we use it

Your recordings and metadata are used to:

  • Train and evaluate Kashmiri speech-to-text models.
  • Improve dataset quality through review by our team.
  • Publish an open Kashmiri speech corpus, where audio may be redistributed alongside the recording metadata above. Names, contact details, and account identifiers are not published with the corpus.

Storage and sharing

  • Mobile app: recordings stay on your device until you choose to upload or export them. When uploaded, audio is sent to our secure servers along with the metadata above.
  • Web recorder: recordings are uploaded directly to our servers when you tap Save. Your session token is stored in your browser's localStorage and is valid for up to 30 days.
  • Hosting: our backend and audio storage are hosted on Railway, with servers located in the United States. Postgres is used for metadata; audio files are stored on an encrypted volume.
  • Retention: recordings and metadata are retained indefinitely as part of the dataset. Profile fields can be edited or cleared from the app at any time.
  • We do not share your name or contact information with third parties. Aggregate, anonymised statistics and the audio dataset may be shared with research partners and published as an open corpus.
  • We do not use third-party analytics or advertising trackers.

Your rights

  • Access: you can view your uploaded recordings in the History section of the web recorder and in Contributions in the mobile app.
  • Export: you can export all your local recordings from the mobile app via Settings → Export recordings.
  • Deletion: to request deletion of recordings or profile data already uploaded to our servers, email farhansyedain@gmail.com from the email address associated with your account. We aim to respond within 14 days.
  • Sign out: signing out removes your active session. On mobile, local recordings remain until you uninstall the app. On the web, signing out clears the session token from your browser.

Permissions

  • Microphone: required to record your voice when you tap Record. In the web recorder you'll be prompted by your browser; in the mobile app, by the system permission dialog.
  • Storage (mobile only): required to save recordings locally before they are exported or synced.
  • Internet: optional on mobile (used to fetch new prompts and sync recordings when you are online); required on the web recorder.

Children

The app is intended for users 18 and older, or for younger users with the consent of a parent or guardian. We do not knowingly collect data from children under 13.

Changes

We may update this policy. The "Last updated" date at the top reflects the most recent revision. Material changes will be communicated through the app.

Consent

By tapping Continue in the onboarding flow, or by recording in the web recorder, you agree to the above and consent to your voice being used to train Kashmiri speech models and to be redistributed as part of an open Kashmiri speech corpus.