mirror of
https://github.com/OpenSquawk/OpenSquawk
synced 2026-08-01 14:16:05 +08:00
Controllers now speak with the four Kokoro speakers (bm_george, bf_emma, am_michael, af_heart — heard most, best quality), pilots with six distinct Piper speakers (alan, alba, joe, amy, bryce, kristin); ATIS moves to the dedicated ljspeech broadcast voice. The classroom instructor draws a stable controller-pool voice per module when the user hasn't configured one, and the TTS cache key follows the effective voice. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>