Voice Consent Policy

Voice features in Ethen today#

How a voice conversation in Ethen Chat works#

Starting a session

Voice starts only when you choose it. When you start a voice conversation, your browser asks for permission to use your microphone. If you allow it, Ethen gives your browser a single-use session credential that expires after ten minutes. Your browser then opens a direct, encrypted connection to the realtime voice service.

Where your audio goes

While the session is active, your browser streams microphone audio directly to the Vercel AI Gateway, which carries it to OpenAI's realtime voice model. The model listens, responds with speech, and also produces text of the conversation. Because your browser connects directly, Upcube's own servers do not receive, relay or record your audio. The voice provider receives your audio and your IP address, and processes them under its terms.

What is saved

When the voice exchange ends, Ethen saves the text of what was said, by you and by Ethen, to your conversation, alongside your typed messages. It is stored, kept and deleted like the rest of the conversation, as described in the Data Retention Policy. Ethen does not save an audio recording of the conversation.

What is not happening

• Ethen does not listen when no voice session is active. There is no background listening or wake word.
• Ethen does not use your voice to identify you, and does not create a voiceprint.
• Ethen does not build a model of your voice from voice conversations.

Ending and deleting

You can end a voice session at any time. You can also revoke Ethen's microphone permission in your browser's site settings. To delete a voice conversation's transcript, delete the conversation. Copies of audio or text the voice provider received during the session are governed by its terms.

Audio in Ethen Studio#

Studio's audio models generate speech and sound from text prompts and settings. Studio sends those prompts to the model's provider, as described in the Ethen Studio Media, Likeness & Generated Content policy. When uploads are turned on, an audio file you upload as a reference will be scanned before acceptance and stored privately in your project. Uploads are currently turned off. Generated audio is stored in your Studio project until you delete it.

Your responsibility for other people's consent#

Your microphone can capture more than your own voice. Voice features are intended for your own speech. If other people may be heard, you are responsible for making sure that capturing and processing their voices is lawful and that they have agreed where agreement is needed.

Before you use voice around others

• Tell people. Let anyone who might be heard know that you are using an AI voice feature, and that what is said will be processed by an AI service and saved as text.
• Get permission where required. Many places require the consent of everyone in a conversation before it is recorded or transcribed. Some require it from every party; others from at least one. Follow the strictest rule that applies to the people involved.
• Respect a refusal. If someone does not want to be captured, stop the voice session or move somewhere they cannot be heard.
• Avoid bystanders. Do not use voice features in places where people nearby have not been told and may not expect it.

Meetings, calls and workplaces

Do not use Ethen's voice features to record or transcribe meetings or calls with other people unless:

• every participant has been told;
• you have obtained any consent the law or your organization requires.

At work, use voice features only where your employer's policies allow it, and never to capture confidential conversations you are not authorized to share with a third-party AI service.

Sensitive conversations

Avoid using voice features for conversations involving:

• health information;
• legal advice;
• financial account details;
• information about children;
• other sensitive matters;

unless you are comfortable with that content being processed by the voice provider and saved as text in your conversation.

Common situations

Accessibility#

Voice features can make Ethen easier to use for people who find typing difficult. If you rely on voice for accessibility, the same processing applies: your audio goes to the realtime voice provider and the transcript is saved to your conversation. If you need an alternative way to use Ethen, you can type instead at any time.

Organizations#

Ethen does not yet offer Organization controls for voice features. When it does, administrators may be able to turn voice off for their members or limit where it can be used. Until then, organizations that do not want their staff to use voice features should set that rule in their own policies.

What Ethen does not verify#

Ethen does not verify who is speaking, whether other people were present, or whether anyone consented to being captured. Starting a voice session is your confirmation that you have met the responsibilities above. Ethen is not in a position to confirm that consent was given.

Children#

Do not use voice features to capture children's voices without the permission of a parent or guardian. The rules for who may use Ethen are in the Terms of Service.

Reporting concerns#

If you believe someone has used Ethen's voice features to capture your voice without consent, report it through the contact options in the Legal & Trust Center. Include as much detail as you can, such as the date and context. Misuse of voice features violates the Acceptable Use Policy.