Fix runtime NameError (get_api_client) and CI ModuleNotFoundError (absolute imports)

Fix Docker build: add pkg-config/ffmpeg-dev, fix typo st-audiorecorder

fix: resolve build errors (undefined variables) and restore missing methods

fix: restore HF config in README and fix undefined variables in stt/sign services

fix: replace unmaintained TTS with coqui-tts fork

fix: remove melotts (broken package) from all requirements

fix: resolve dependency conflicts for Hugging Face (TTS/MeloTTS)

feat: Phase 18 Universal Hub, S2S Bridge, and Sign LSTM Simulation

Initial commit: Complete VoiceForge Enterprise Speech AI Platform