Show HN: Multi-modal Muse assistant on an ESP32

Category: ai-ml

Tags: esp32, voice-assistant, embedded, openai-compatible, streaming, iot

Score: 7.3/10 (Innovation: 7, Technical: 8, Documentation: 8, Utility: 6)

esp32-thinks is a firmware project that turns an ESP32-S3 (SenseCAP Watcher) into a voice assistant with wake word, streaming speech-to-text, chat, and text-to-speech, all via configurable OpenAI-compatible endpoints. It stands out for its provider-agnostic design, on-device streaming and barge-in, and an optional local agent for memory and tools, making private, customizable voice AI accessible on cheap hardware.

Target audience: embedded developers, IoT enthusiasts, AI hobbyists

Repository: https://github.com/raullenchai/esp32-thinks · Shell · MIT · 244 stars

View on Hacker News