Your workspace. Your way.
Tune the voices, set the mood, and know where your data goes.
Voice & conversation
Use your browser’s text-to-speech for AI answers.
After speaking, listen for your next question. Starts only when you tap the microphone; stops when you leave or hide the assistant screen. Browser support varies.
Current match: System default; voices may load after speech starts.
Current match: System default; voices may load after speech starts.
Voice gender is not exposed by the browser. Automatic matching uses voice names; pitch alone cannot guarantee a male or female sound. Test and select two distinct voices above. Android may provide only one English voice—additional free voices depend on your device’s speech engine.
Appearance
Neon accents and a little less glare.
Privacy, without the fine print
AI requests: Your messages, earlier messages in the current conversation, and the selected assistant’s saved memories are sent to the external AI service through GenMB. Answers are AI-generated and may be incorrect; they are not live web results. Do not share secrets or highly sensitive information.
Microphone & speech: Permission is requested only after tapping the microphone. Browser speech recognition may transmit audio to its provider and need a network connection. Online speech voices may send reply text to the voice provider. This app does not store recordings. It stops listening when hidden or when you leave the assistant screen.
Saved content: Chats and memories use AES-GCM encryption before being saved in the app’s built-in key-value store. Only this browser holds the encryption key. No sign-in capability is enabled, so there is no account-based storage or sync. The anonymous storage namespace is shared-server storage, not an authenticated account: encryption protects content, but does not prevent someone with the storage key from deleting it. Clearing browser data loses access to old history.
No phone control: No contacts, private files, calls, messages, app launching, or background recording. This is a conversational assistant, not a device controller.
Voice preferences and appearance are saved locally. Memory is explicit: only information you add in the Memory bank is remembered between chats.
Free AI & honest limits
No API key or paid credits required for basic testing. The enabled GenMB AI service provides 100 requests per hour on the Free plan, shared across all visitors to this app. Responses can take up to 60 seconds. This app requests up to 1,800 output tokens per answer. Rate-limit and connection errors appear in the chat with a retry control.
Not unlimited, not offline: AI requires an internet connection. There is no live web search on this plan. The key-value store allows 600 reads and 180 writes per minute shared across this app; long conversations have a storage size limit. Browser speech has its own availability and provider conditions.
Use it on Android
Publish this app and open its published HTTPS link in Chrome on Android. Choose JARVIS or FRIDAY, type a question, or tap the microphone and allow permission. Use the speaker icon on any reply to read it again. Test and select your preferred voices above.
This is a mobile web app, not a native APK. Chrome’s “Add to Home screen” can create a browser shortcut; the app still needs connectivity. If the embedded preview blocks the microphone, open the published app directly.
Developer setup: Run npm install, then npm run dev. Production: npm run build. The built-in AI and storage SDK is provided by the GenMB preview and published hosting; a standalone Vite server cannot supply it. There are no secrets to configure.
Why do both assistants sound alike?
Some devices expose just one voice. Choose different voices if listed and use each Test button. The app cannot install voices or guarantee gender when the system does not offer distinct voices.