Recording & AI training.

We capture and store your voice to provide this spoken language learning service.

Notice version: 2026-10-10.v4-audio-service

What you are opting in to

We capture and store your voice to provide this spoken language learning service. Before connecting to an audio lesson, you agree to the audio collection, sharing, and use described in the notice. This includes your microphone audio, your voice, accent, pronunciation, and background sounds. We store audio for quality assurance and to improve the service for everyone, including training speech and noise cancellation models using original recordings.

Original audio may retain characteristics that identify you. We do not promise to anonymize, alter, or synthesize your voice before training. Avoid sharing names, addresses, contact details, or other private information during a lesson.

Anonymous learning and recordings

You may learn without creating an account. “Anonymous learning” means an account and email address are not required. It does not mean that your recorded voice is anonymous or impossible to identify. Guest progress uses a browser session identifier. Account details and training recordings are stored separately, but session identifiers link recordings to consent records.

What this app stores

• A record of the notice you accepted and when you accepted it.
• Your session history and learning progress.
• If you create an account: your name, email address, and a salted password hash.

Your microphone audio travels over WebRTC through LiveKit to the separately hosted lesson agent. The agent is responsible for recording and storage; this website does not record or upload audio files. Browser and agent audio processing may affect the recorded audio. We use third-party AI services to provide this service. Your audio and/or text from your conversation is transmitted to those providers to deliver your lesson. Those providers process the data under their applicable terms and privacy policies.

Your choice before recording

You can decline and browse the hub without microphone access. Starting an audio lesson requires opt-in. While connected, the app shows your audio connection status. End the lesson at any time to stop sending microphone audio. The app does not provide a way for users to access or delete their recordings. Ending a lesson stops new microphone capture; recordings already collected may continue to be retained and used for AI training.

The website stores consent and session metadata and sends signed consent claims to the lesson agent. It does not store recordings, export datasets, or run model training. Recording and storage are managed by the separately hosted lesson service. This notice describes the controls currently available in the app; it does not limit any rights you may have under applicable law.

Before a public launch

This is a development implementation. Recording retention periods, model training workflows, service-provider data handling, account recovery, and a production privacy policy must be defined before inviting public participants.