Last updated: 3 June 2026
This policy describes what TalkPass ("we", "us") collects when you use the TalkPass mobile app and Progressive Web App (collectively, the "Service"), and what we do with that information. We try to keep this short and concrete — if we don't mention a data flow on this page, we don't do it.
When you press the talk button, your microphone audio is captured and sent to our server for speech-to-text. The audio is processed in real time and not retained after transcription completes — typically within seconds. We do not record sessions to disk.
The transcribed text of each turn, plus its translation, is shown on your screen. For shared sessions, it is also relayed to your conversation partner. By default, transcripts are deleted within minutes of a session ending. If you explicitly tap "Save" at the end of a session, the conversation is retained in your private history for up to 1 month, then automatically deleted. You can delete a saved conversation manually at any time.
The web (PWA) version never saves conversations.
If you opt into the "My Voice" feature, a 30–90 second audio sample of your speech is sent to our voice-cloning providers (see section 2) to create a synthetic voice model. The raw sample is deleted from our servers within 24 hours; the resulting voice model is kept with the provider until you delete your account or re-record. You can re-record or remove your voice clone in Settings.
When you sign in with Apple or Google, we receive your email address, name (if you chose to share it), and a unique account identifier. We store this in our authentication database (Supabase) to maintain your session and link your subscription / saved conversations to your account.
Your subscription tier (Free / Standard / Pro) and remaining usage minutes are stored in our database. Purchase details (card numbers, billing addresses) are handled entirely by Apple or Google — we never see them.
If a crash or error occurs, we log the technical details (error message, timestamp, anonymised user id) for up to 30 days to debug. These logs never contain your audio, transcripts, or personal content.
To deliver real-time translation we route data through specialist providers. Each receives only the data needed for its function and is contractually required to delete that data after processing (where applicable):
| Provider | What they process | Why |
|---|---|---|
| Supabase | Account info, sessions, saved conversations | Authentication and database hosting |
| Apple / Google Sign-In | Authentication tokens | Sign-in identity |
| RevenueCat | Subscription state and purchase events | Subscription management |
| Groq (Whisper) | Your spoken audio | Speech-to-text transcription |
| Deepgram | Your spoken audio | Alternative speech-to-text |
| Google Cloud Translation | Transcribed text | Default translation |
| Groq (LLaMA 3.1) | Transcribed text | Context-aware AI translation |
| OpenAI | Transcribed text | Fallback for AI translation |
| Google Cloud Text-to-Speech | Translated text | Default synthetic voice (Neural2) |
| Microsoft Edge TTS | Translated text | Fallback synthetic voice |
| Cartesia | Voice samples, translated text | Voice cloning (Pro) |
| Fish Audio | Voice samples, translated text | Voice cloning (Pro) |
| Fly.io | All server traffic | Application hosting |
We do not sell your data to anyone, and we do not share it with advertisers or analytics networks.
You can:
If you are in the EU/EEA, UK, or California, you have additional rights under GDPR and CCPA respectively, including the right to object to processing and to lodge a complaint with a supervisory authority. Email us at the address above to exercise these rights.
You may only create a voice clone of your own voice. By recording the sample, you confirm that the voice belongs to you and that you authorise TalkPass to generate a synthetic model from it for use within the app. Creating a clone of someone else's voice without their explicit consent is prohibited and violates our Terms of Service.
TalkPass is not directed at children under 13. We do not knowingly collect data from users under 13. If you believe a child has provided us with personal information, contact us and we will delete it.
Traffic between the app and our servers is encrypted with TLS. Account data is stored on Supabase (PostgreSQL with row-level security). We do not store your audio in any form after transcription. No security system is perfect — if we discover a breach affecting your account, we will notify you within 72 hours.
If we make material changes to this policy, we will update the "Last updated" date at the top and, for substantive changes affecting how your data is processed, notify you in-app before they take effect.
Questions or requests: support@talkpass.app.