Last updated September 22, 2026
This policy explains what A6 Relay ("we," "us") collects when you use the site or the interpretation product, and why. A6 Relay is currently in active testing and early access — this policy describes what the product actually does today, and we'll update it as it changes.
Speech recognition and speech synthesis both run entirely inside your own browser, using on-device machine learning models. When you speak during a session, that audio is processed locally on your device and is never transmitted to, or stored by, A6 Relay's servers or infrastructure — not even temporarily. The only thing that ever leaves your browser is the resulting text, described below.
Your browser does download the speech-recognition and speech-synthesis model files directly from our model provider's (Moonshine's) content delivery network the first time you use a session. These are generic model files, not anything specific to you or your conversation, and no audio or personal information is sent as part of that download.
Once your speech is transcribed in your browser, the transcribed text and its translation are sent to our servers to produce and display the translation, and are stored so a session's transcript can be displayed, so trial usage limits can be enforced, and so our team can review translation quality and improve accuracy over time (including flagging and correcting mistranslations). This applies to every session, whether or not you've created an account.
To produce a translation, the transcribed text (not audio) is sent to our translation provider, OpenRouter, which routes it to one of several third-party AI language models. That text is processed by whichever model is handling the request in order to return a translation, subject to that provider's own data handling practices.
If you create an account and choose to save a conversation, that transcript is stored against your account so you can view and edit it later. Trial sessions (used without signing in) are not saved to an account, but the underlying transcript text is still retained on our servers as described above.
If you create an account, we collect the name and email address you provide. Authentication (including password handling) is managed by our database provider's managed authentication service — we don't store your password in plain text. Signing in sets a session cookie so you stay signed in between visits; we don't use advertising or third-party tracking cookies.
If you join a waitlist or send us a message through our contact form, we collect the email address (and, for contact messages, the name and message content) you provide, so we can follow up with you. This information is stored in our database and in a private, internal spreadsheet — it is not published or made public, and we don't sell or share it with third parties for their own marketing purposes.
We use a small number of service providers to operate A6 Relay: a database and authentication provider to store account, session, and transcript data; a translation routing provider (and the underlying AI model providers it connects to) to produce translations; and standard web hosting. We don't sell your personal information, and we only share it with these providers as needed to operate the service.
We retain transcript, account, and waitlist data for as long as needed to operate and improve the service, unless you ask us to delete it sooner. To request deletion of your account or the data associated with it, or to ask what data we hold about you, reach out through our contact page.
A6 Relay is not directed at children, and we don't knowingly collect information from children under 13. If you believe a child has provided us information, contact us and we'll remove it.
As A6 Relay moves from early access toward a full launch, this policy will be updated to reflect how the product actually works at the time. We'll update the date at the top of this page when that happens.
Questions about this policy or your data can be sent through our contact page.