I wanted to explore what an AI assistant would feel like if it was designed as a product instead of just another chatbot.

So I built JISSI.

JISSI is a cross-platform AI voice assistant built with React Native, Expo, React Native Skia, and OpenRouter.

Instead of focusing only on conversations, I designed it around three principles:

• Natural voice interactions

• Beautiful, expressive UI

• Extensible AI architecture

Current capabilities include:

• Real-time speech recognition

• Natural text-to-speech

• Context-aware conversations

• OpenRouter-powered LLM responses

• Animated GPU-powered energy orb

• Cross-platform support (Android + Web)

• Modular tool architecture for future integrations

• Modern luxury-inspired interface

The long-term goal is to evolve JISSI into an assistant that can interact with tools, understand context, and feel genuinely conversational instead of robotic.

Still actively improving:

• Lower voice latency

• Better multilingual support

• Tool calling

• Calendar & weather integrations

• Memory improvements

I'd love feedback from builders, designers, and AI engineers.