Inflect-Micro-v2: Apache-2.0 text-to-speech in 9.36M parameters at 6.28x real-time on four CPU threads.

owensong's VITS-family English TTS has 9,356,513 deployable parameters and a 37.53MB FP32 footprint producing 24kHz mono (Hugging Face). Reported: 66.2% preference in community blind listening tests, 4.395 UTMOS22 naturalness, 3.99% semantic error rate across multiple ASR evaluators. It hit 180 points on HN with only 15 comments, which usually means people upvoted and went to try it. The training corpus pipeline and full optimization recipe are withheld, so it's a usable artifact rather than a reproducible one.