ElatoAI
Realtime Voice AI with 100+ Models on Arduino ESP32 with Secure Websockets & Edge Functions for AI Toys, Desk Companions and Devices
What is ElatoAI?
Realtime Voice AI with 100+ Models on Arduino ESP32 with Secure Websockets & Edge Functions for AI Toys, Desk Companions and Devices
Overview of ElatoAI
What it is
ElatoAI is an open-source project that provides realtime voice AI capabilities on Arduino ESP32 devices. It supports over 100 voice AI models and uses secure WebSockets and edge functions to enable uninterrupted conversations globally. The project is designed for AI toys, desk companions, and other devices.
Key features
- Realtime speech-to-speech conversion using APIs from OpenAI, Gemini, xAI, Eleven Labs, Boson Higgs, and AssemblyAI.
- Support for 100+ STT, LLM, and TTS models via FastAPI server or Cloudflare Workers.
- Secure WebSocket communication with Opus audio compression.
- Server VAD turn detection for smooth conversation flow.
- Global edge performance with Deno Edge Functions or Cloudflare Workers.
- ESP32 Arduino framework integration with no PSRAM required.
- OTA updates, WiFi management with captive portal, and factory reset from a Next.js webapp.
Use cases
ElatoAI can be used to create AI-powered toys, desk companions, and other devices that require realtime voice interaction. It is suitable for building custom AI agents with different personalities and voices, and for deploying multiple devices globally.
Getting started
See the repository README for detailed instructions on quick start, building with PlatformIO, building on Arduino IDE, and deploying globally.
License and pricing
The license is not stated. The project is listed as open source.
Based on the project's public README and repository metadata.