russell@unturf.com c3ac2def48 Add idempotent voice registry system for permanent speaker-to-name assignments ... - voice_registry.json: append-only registry with 50 name pools per gender, locked speaker assignments, and multi-corpus support - Rewrite download script to be registry-driven: loads registry, assigns names deterministically (sorted by speaker ID), never changes existing assignments - Update docs/VOICES.md with registry system documentation - Support --registry and --corpora CLI flags for multi-corpus downloads		2026-01-27 13:59:39 -05:00
..
models	Standardize project naming to uncloseai-speech across all files	2025-11-10 05:23:34 -05:00
research	Standardize project naming to uncloseai-speech across all files	2025-11-10 05:23:34 -05:00
AUDIT.md	Standardize project naming to uncloseai-speech across all files	2025-11-10 05:23:34 -05:00
MIRRORS.md	Standardize project naming to uncloseai-speech across all files	2025-11-10 05:23:34 -05:00
MODELS.md	Make Qwen3-TTS the default engine, add CPU-only docker support	2026-01-26 10:41:23 -05:00
VOICES.md	Add idempotent voice registry system for permanent speaker-to-name assignments	2026-01-27 13:59:39 -05:00