Complete local text-to-waveform speech synthesis under 10M parameters

Inflect-Micro-v2, a complete local text-to-waveform speech synthesis model with under 10M parameters, has been released, offering fixed-voice English TTS with deterministic seeds, long-text handling, and CPU or CUDA inference. The model is part of Inflect v2, which provides a single public API across two sizes: Micro (under 10M parameters) and Nano (under 4M parameters).