A new perspective on your everyday questions.
LOCAL AI · OLLAMA
Free local AI · uses your installed Qwen2.5 3B model. Larger models need more memory and a one-time download.
Connect this phone to Ollama on your computer (not installed on the phone):
npm run dev.ollama pull command on the computer.This works when the app is served from the computer running Vite. It cannot install Ollama or change device settings. Use a trusted Wi-Fi network.
Ollama and LM Studio run locally without API keys, but require their server on the same computer as this app. A public static website cannot connect to a local server; use Demo replies online. Speed depends on your model and hardware. Microphone permission is requested only when you start voice input.