kokoallovero

kokoallovero is a statically compiled, fully self-contained text-to-speech inference engine using the Kokoro-82M model optimized for CPU deployment. It includes an English-only grapheme-to-phoneme system, text normalization, and SIMD acceleration for high performance without dependencies on Python or runtime downloads. Outputs 24 kHz audio from input text with multi-threaded synthesis, suitable for portable deployment on moderate CPUs.