Settings & privacy.
Tune your assistants’ voices, understand your free plan, and stay in control of your data.
Voice preferences
Voice gender isn’t reliably reported by browsers. We prefer known masculine/feminine voices and subtly adjust pitch, but cannot guarantee gender or availability. Test and choose a voice that sounds right to you. Android voice availability depends on your installed speech engine.
Conversation preferences
Hands-free mode starts only after you tap the microphone. Use Stop, turn off voice mode, leave the chat, or background this page to stop listening. Some browsers require a new tap for every question. Speech recognition is detected in this browser.
Privacy, by design
Messages go to an external AI service. The app sends your question, up to 30 recent messages from that conversation, and the memories you assigned to that assistant through the platform’s built-in AI. Responses are AI-generated, not live-searched. Provider retention policies are controlled by the platform, not this app.
Your microphone is opt-in. It is requested only when you tap the mic. Browser speech recognition may send audio to its speech provider. This app sends the resulting text—not audio—to the AI, and does not store audio. A non-local speech voice may send response text to its voice provider.
Your vault is encrypted. Chats and memories are AES-GCM encrypted in your browser before saving to the app’s built-in key-value store. Only the random vault identifier and decryption key stay on this device. No plaintext private chats or memories are written to shared storage.
Know the limits. There is no account-based sync. Clearing site data loses your key and access to saved history. Shared storage does not provide account ownership or protection against someone with the vault identifier deleting ciphertext. Encryption protects contents, not service availability. Don’t use this app as your only copy of important information.
No phone control. Neither assistant can open other apps, send messages, make calls, inspect files, or read contacts.
Free plan & getting started
No AI key or paid credits required to try. The enabled built-in AI allows 100 requests per hour shared across this app on the Free plan. Errors show in chat when limits are reached. Requests may take up to 60 seconds. We cap each answer at 1,800 output tokens. Free is limited, not unlimited; live web search is not included.
How do I use this on Android?
Publish the app with the platform’s Publish control, then open its published HTTPS URL in Chrome on Android. Pick an assistant and type a question. For voice, tap the microphone and grant permission. You can use Chrome’s Add to Home screen option for a shortcut. This is a mobile web app, not a generated APK. Microphone access may be blocked inside the editor preview, so test on the published page.
How do I get different voices?
Choose an installed voice from the selectors and test both assistants. If needed, install or change your Android text-to-speech engine in system settings, then reopen the browser. Browser/system voices are used without a separate paid voice API. The app cannot install a voice itself.
What should I verify on my phone?
Send a typed question, then a follow-up. Switch assistants and confirm separate histories. Reopen and rename a chat. Add and delete a memory. Test both voices, speak into the microphone, interrupt a response, deny microphone access, and test hands-free follow-ups. Voice support depends on your actual device; we cannot certify microphone or audible voice output from the code preview.