Show HN: Babytalk: Offline speech to text and text to speech on ESP32

Category: ai-ml

Tags: esp32, speech-recognition, text-to-speech, embedded, edge-ai

Score: 8.3/10 (Innovation: 9, Technical: 9, Documentation: 8, Utility: 7)

BabyTalk enables offline, general-purpose English speech-to-text and text-to-speech directly on ESP32-S3/P4 microcontrollers, using quantized neural models stored in flash and run with tight RAM budgets. It stands out by bringing cloud-free, command-free voice interaction to embedded devices like LoRa radios and screenless nodes, with bindings for C, MicroPython, and Erlang/Elixir.

Target audience: embedded developers, IoT engineers, makers

Repository: https://github.com/tlack/babytalk · C · NOASSERTION · 47 stars

View on Hacker News