Effective: April 24, 2026 · Last updated: April 24, 2026
This page describes privacy practices specifically for LLT Mobile, the iOS companion app for Local Live Translator. For the general adelvo.io website policy and GDPR/CCPA rights, see the general Privacy Policy.
LLT Mobile captures audio from the device microphone, sends it to a user-selected speech-to-text engine for transcription and translation, and displays the result on screen. Optionally, translations are forwarded to a local vMix server on the user's Wi-Fi, or audio is streamed to the user's own Mac running the LLT desktop app for processing there.
No server, no account, no tracking. Adelvo does not operate speech-to-text infrastructure and does not receive, route, or store any audio or transcripts.
For the full legal entity details, see the Impressum.
LLT Mobile uses the iOS microphone exclusively while the user has an active translation session running. Audio is processed in real time and streamed to the selected engine. No recording is persisted on Adelvo's servers at any time — Adelvo does not operate speech-to-text infrastructure and has no backend that receives audio.
iOS requires explicit microphone permission on first use. You can revoke it at any time in Settings → Privacy & Security → Microphone.
The user chooses where audio is sent. There are two operating modes:
The user provides their own API key. The resulting data handling is governed entirely by the selected provider's privacy policy. Adelvo does not intercept, copy, or route this data through its own infrastructure.
| Engine | Destination | Provider Privacy Policy |
|---|---|---|
| OpenAI Whisper API | OpenAI servers | openai.com/policies/privacy-policy |
| Deepgram | Deepgram servers | deepgram.com/privacy |
| Google Cloud Speech | Google servers | cloud.google.com/terms/cloud-privacy-notice |
| Microsoft Azure Speech | Microsoft servers | privacy.microsoft.com |
When the user selects LLT Backend, audio is streamed from the iOS device directly to the user's own Mac running the LLT desktop app, over the local Wi-Fi network. The stream never leaves the user's LAN on its way to the Mac.
The Mac then processes the audio according to the engine configured in the desktop app's settings. Depending on that choice, audio may either stay on the user's Mac or be forwarded from the Mac to a cloud provider — always under the user's own API key:
| Engine (on Mac) | Final Destination | Provider Privacy Policy |
|---|---|---|
| Whisper (local, MLX) | Stays on the user's Mac | — fully local, no cloud |
| Vosk (local, self-hosted) | Stays on the user's Mac | — fully local, no cloud |
| OpenAI Whisper API | OpenAI servers | openai.com/policies/privacy-policy |
| Deepgram | Deepgram servers | deepgram.com/privacy |
| AssemblyAI | AssemblyAI servers | assemblyai.com/legal/privacy-policy |
| Google Cloud Speech | Google servers | cloud.google.com/terms/cloud-privacy-notice |
| Microsoft Azure Speech | Microsoft servers | privacy.microsoft.com |
Regardless of mode or engine, Adelvo has no server in between and receives no audio, transcripts, API keys, or usage telemetry from either app.
If the user enables transcript saving in Settings, text files (TXT and SRT) are written to the app's local storage on the device. These files are accessible only to the user via the Files app or the in-app transcript list, and may be shared via the standard iOS share sheet (AirDrop, Mail, Messages, etc.). They are never uploaded anywhere by LLT Mobile.
If enabled, translated text is sent via HTTP directly from the device to a user-configured vMix server IP and port on the local network. This traffic never leaves the user's Wi-Fi.
If enabled, LLT Mobile can speak the translated text aloud using Apple's on-device speech synthesizer (AVSpeechSynthesizer). The synthesized audio is routed to the active iOS audio output and is intended to be played through connected headphones — this is by design, to prevent the spoken translation from being picked up by the microphone and creating a feedback loop.
No text is sent to any cloud text-to-speech service. Speech synthesis happens entirely on the device; no data leaves the iPhone/iPad for this feature.
API keys for cloud engines are stored only on the device, in iOS UserDefaults. They are never transmitted to Adelvo.
LLT Mobile Pro is sold via Apple's in-app purchase system. Payment is handled entirely by Apple. Adelvo receives no payment details, credit card information, or personal billing data — only an anonymous entitlement status from Apple's receipt validation.
LLT Mobile contains no analytics SDKs, no tracking, no advertising identifiers. No telemetry is sent to Adelvo or any third party from the app itself.
Because LLT Mobile does not send any personal data to Adelvo, there is typically no data on our side to access, rectify, or delete. If you nevertheless have a request, contact us at hello@adelvo.io. Your full GDPR and CCPA rights are described in the general Privacy Policy.
For data held by the cloud speech providers you choose to use with your own API key, please exercise your rights directly with the respective provider.
LLT Mobile is not directed at children under 13 and does not knowingly collect data from them.
We may update this policy as the app evolves. Changes will be posted on this page with an updated effective date.
For any privacy-related questions or requests regarding LLT Mobile: