Voice & AI
How Audiofirst uses AI
Read what the in-app AI disclosure says stays on your phone and which three features send data to a server or an AI service.
What it does
Audiofirst has an in-app AI (artificial intelligence) disclosure called How Audiofirst Uses AI. It says what stays on your device and what leaves it. Text recognition from screenshots, photos and the camera, flashcard generation and speech recognition run on your device. Three features send data off your device: premium voices, Live Voice Tutor and Read formulas. Some pack content was written or translated with AI assistance before it was published.
How to use it
- Finish onboarding on a new install. The AI-Generated Content screen opens if you have not seen it yet.
- Read the five points on the screen.
- Tap Learn more to open the full text, How Audiofirst Uses AI.
- Tap I Understand to close the screen. You cannot close it by swiping.
- To read the full text later, open Settings and scroll to Data & Privacy.
- Tap How Audiofirst Uses AI.
- To see the first-launch screen again, tap View AI Disclosure in the same section.
Settings that affect it
- Settings → Data & Privacy → How Audiofirst Uses AI — opens the full text. It is always available.
- Settings → Data & Privacy → View AI Disclosure — shows the AI-Generated Content screen again at once. It has no on or off state.
Limits and honest caveats
- The full text says three features send data off your device.
- Premium voices: when you play a card, its text goes to our server and a third-party text-to-speech provider (ElevenLabs, Microsoft Azure, OpenAI or Google, depending on availability). Audiofirst falls back to the on-device voice if that is unavailable. See cloud voices.
- Live Voice Tutor: when you start a session, your microphone audio is streamed to our server and a third-party realtime AI provider for that session only. See Live Voice Tutor.
- Read formulas, where the feature is available: the one photo you choose goes to our server and to OpenAI to read the formulas in it. The app asks you to confirm every time. The disclosure says Audiofirst does not keep the photo.
- The two session features start only when you start them. Premium voices work when you play a card.
- The short first-launch screen is a summary. It does not mention Read formulas. The full text does.
- The full text can be updated from our server without an app update. The app shows the bundled text first, keeps a saved copy, then refreshes it when the network allows.
- Cards written by an AI model carry an AI-generated badge under the revealed answer. The first-launch screen says content is based on official sources but may contain inaccuracies, so check critical facts yourself.
- This page describes the disclosure. It is not a privacy policy and not legal advice.
FAQ
Does Audiofirst send my screenshots or photos to an AI service?
Not for text recognition. The disclosure says recognition from screenshots, photos and the camera runs on your device. The exception is Read formulas, where you confirm each photo before it is sent.
Is my speech sent anywhere when I answer by voice?
Speech recognition runs on your device, according to the disclosure. Microphone audio is streamed off your device only in a Live Voice Tutor session that you start.
Can I read the disclosure again?
Yes. Open Settings, then Data & Privacy, then How Audiofirst Uses AI.
Why did the screen appear again?
It appears when you tap View AI Disclosure in Settings. It also appears after onboarding if you have not tapped I Understand yet.