Add TypecastEngine

[jaebong-human]

Summary

• Adds TypecastEngine using the official typecast-python SDK
• Adds TypecastVoice with voice metadata (name, voice_id, gender, age, use_cases, models)
• Registers both in the lazy-loading __init__.py files

Details

Constructor parameters:

• voice_id — Typecast voice ID (or TYPECAST_VOICE_ID env var)
• model — ssfm-v30 (default) or ssfm-v21
• emotion_type — "preset" or "smart" (ssfm-v30 context-aware inference)
• emotion_preset — normal, happy, sad, angry, whisper, toneup, tonedown
• emotion_intensity — 0.0–2.0
• tempo, pitch, volume, language, seed
• api_key (or TYPECAST_API_KEY env var — required)
• host — optional API endpoint override (SDK also reads TYPECAST_API_HOST)

Audio format: WAV, mono, 44100 Hz, 16-bit PCM

Installation:

pip install typecast-python

Usage:

from RealtimeTTS import TextToAudioStream, TypecastEngine

engine = TypecastEngine(
    voice_id="tc_685c9a70f12503316e5359d2",
    emotion_preset="happy",
)
stream = TextToAudioStream(engine)
stream.feed("Hello from Typecast!")
stream.play()