Files are read in this browser and sent to your selected AI provider when you press send.
Microphone: checking…
Speech recognition: · Speech output:
AI backend:
About JARVIS — a voice assistant by Arifin Wahid. This web version runs in the browser; device control (flashlight, Wi-Fi, volume, wake word) requires the native Android version.
Optional. Add your own API key for full conversations. The key is saved only in this browser and sent only to the provider you pick. Don't use this on a shared device. Keys for several providers can be saved and switched from the top bar. A Custom API receives your key, so only use one you trust. If you skip, JARVIS still handles many commands offline.
JARVIS Voice Assistant, version 2. Made by Arifin Wahid.
No account is needed and this website has no server that stores your data.
Your chat history, settings, notes, tasks and API key are saved only in this browser on this device. Delete them anytime with "Clear chat history" and "Remove saved key". On a shared device, always remove your key.
Voice input: Chrome sends your speech to its own speech service to turn it into text, under that service's privacy policy.
AI questions go directly from your browser to the provider you choose (Groq, Google, OpenAI or Anthropic) using your own key, under that provider's terms. Weather sends your location (only if you allow it) to Open-Meteo. Offline answers to "what is…" questions send your question to Wikipedia. Translation and search open Google in a new tab. Image generation sends your description to the free Pollinations service. When you press send, attached images and extracted text from supported documents are sent to the AI provider you selected. PDF and Office files are read locally in this browser; their reader libraries are loaded from public CDNs on demand, and the original documents are not uploaded to a JARVIS file server. Gemini web search sends your question to Google. A Custom API or backend URL receives your questions (and, for Custom API, your key). Wake-word listening keeps the microphone active while the page is open and uses Chrome's speech service.
Pick a tool, then type or speak in the chat. Use 📎 to attach images, PDFs, Word documents, spreadsheets, presentations or text files. Files are processed in your browser; most attachments need an AI key. Scanned PDFs may not contain readable text.