Privacy Policy

Last updated: 24 September 2026

TalkDash is a speaking-practice app operated by Simpumind. This policy describes what the current product actually does.

What we collect

On your device. Practice progress, reminders you chose, and recordings you keep can live locally on the phone.

Install identifier. The app creates an install id so your device can talk to our servers. It is not an advertising id.

Voice recordings. When you practise with the microphone on, TalkDash records your turns so you can hear them back at the end of the session. Recognising what you said happens on your phone, and your recordings stay on your phone: the app does not upload them. At the end of your second session you are asked whether to keep recordings. Until you answer, the lines from your first session wait on the phone: if you keep recordings they join the ones you keep, otherwise they are deleted, and they are deleted after 7 days if no second session reaches its end. Recordings you keep stay on the phone until you delete them (Settings, Delete all recordings). If you continue without a microphone, nothing is recorded.

Transcripts and notes derived from a recording. These are product data (so you can replay, compare, and return to a line). They are not sold, and they are not used as advertising profiles.

Named analytics events. When product analytics is on, we send a closed list of event names such as finishing a first recording. Those events do not include transcripts, audio files, or the words you said.

Advertising measurement. After you keep a first recording, iOS may ask whether TalkDash can track you across other companies’ apps and websites. That prompt is optional and is not required to practice. If you allow it, and if advertising keys are configured, AppsFlyer and TikTok may receive an advertising identifier and the same named events, so we can tell whether an ad led to this install. If you deny tracking, practice still works; those vendors may still start in a non-tracking mode (device identifier without IDFA, or SKAdNetwork). We do not send recordings, transcripts, or the words you said to advertising networks.

Crash reports. When crash reporting is on, we receive technical crash data. Request bodies, audio URLs, and transcripts are stripped.

We do not collect precise location, contacts, or a public social graph. We do not sell voice data.

Why we use it

• Run the practice you asked for (record, play back, score, remind you on the days you chose).
• Keep an install-scoped account of that practice on our servers when the app is online.
• Understand whether the product is working, using named events, not recordings.
• Measure paid ads when you allow tracking, using an advertising identifier and named events, not recordings.
• Find crashes.

Local notifications exist only for planned practice. The system prompt is asked when you set a reminder, not at first launch.

How long we keep it

Raw uploaded audio is deleted from object storage after 90 days by default (RAW_AUDIO_TTL_DAYS). Derived transcripts and practice history can remain until you delete your data. You can export or erase everything in Settings → Privacy. Erasure calls DELETE /v1/me, removes server rows that belong to the install, and deletes stored audio. If an audio object cannot be deleted immediately, a retry row remains until a later sweep finishes it — the install is still gone, but we do not claim the object is gone until that retry succeeds. Scenario clips that other learners also use are not personal data and stay.

Accounts

You can use TalkDash as a guest. Guest progress stays on that device until you sign in. Apple or Google sign-in, when it is live, stores a provider identifier so you can recover the same account on another device. We do not offer email-and-password accounts.

Processors

Analytics may go to PostHog in the EU when a project key is configured. Crash reports may go to Sentry or a self-hosted GlitchTip when a DSN is configured. Advertising measurement may go to AppsFlyer and TikTok when those keys are configured and, on iOS, after the tracking prompt. Language-pack audio (the lessons' recorded voices) is served from our object storage; your own recordings are not stored there. We do not send recordings to advertising networks.

Children

TalkDash is not directed at children under 13.

Changes

If this policy changes in a way that affects what we keep, we will update the date above. The in-app Privacy link opens this page.

Contact

Write to privacy@talkdash.app for anything about your data, or use Settings → Privacy in the app to export or delete it.