Give intelligent software a voice.
An extensible, voice-first AI platform connecting natural conversations, agent capabilities, and smart hardware.
AI Voice Agent
A modular voice-agent system built around a Rust server, real-time audio, language models and connected devices. The goal is a dependable foundation for assistants that can listen, think, speak and interact with the physical world.
Useful AI should be accessible beyond a chat window. Voice interfaces need low-latency audio, predictable session behavior, flexible providers and safe device interactions.
Develop a composable system with dedicated voice processing, agent orchestration, configurable templates and tool integrations; pair it with ESP32 clients and a management interface.
Core capabilities
Real-time voice
Streaming speech input and output with structured session management.
AI orchestration
LLM conversations with template and provider selection.
Tool ecosystem
Device capabilities and external MCP integrations.
Connected hardware
ESP32-based endpoints for physical-world experiences.
Build roadmap
Voice foundations
Server, audio transport and voice pipeline.
Experience and reliability
Streamlining operations, enrollment and observability.
Product ecosystem
Reusable agent experiences across more devices.