🎙 Local TTS on Jetson Orin Nano via StreamTTS
The StreamTTS architecture has been introduced to run the Kokoro-82M model on the NVIDIA Jetson Orin Nano. Instead of standard REST APIs, the system uses "durable streams" via S2-Lite, turning audio into a single continuous data stream on disk.
🌍 The shift toward a log-centric architecture in AI inference reduces latency and infrastructure complexity.
👤 This serves as an example of creating a reliable voice synthesis service on a budget edge device without relying on heavy cloud solutions.
Source 1: https://s2.dev/blog/local-ai Source 2: https://github.com/s2-streamstore/streamtts