🎭 DramaBox

Expressive TTS with Voice Cloning
🏗️  Built on LTX-2 by Lightricks. DramaBox is Resemble AI's expressive TTS, trained on top of the LTX-2.3 audio branch under the LTX-2 Community License. Huge thanks to the Lightricks team for open-sourcing the base.
🪄 Voice Generator
Scene Prompt Put speech in "quotes", directions outside
Voice Reference (Optional) 10+ seconds recommended
📤
Click to upload or drag & drop
Supports MP3, WAV, M4A, etc.
CFG Scale 2.5
STG Scale 1.5
Duration Multiplier 1.10
Target Duration (s) — 0 = Auto 0.0
Reference Duration (s) 10.0
Generation Seed
🔊 Output Room
Connecting to DramaBox engine...
🎛️

Enter a prompt and hit generate to voice your script

🎬 Quick Demos
📖 Prompt Writing Guide
Structure Pattern

<speaker description>, "<dialogue>" <action> "<more dialogue>"

Inside Double Quotes (Spoken)
  • Standard speech dialogue: "Hello, how are you today?"
  • Phonetic sounds and laughs: "Hahaha", "Hehehe", "Mmmmm", "Ugh", "Argh"
Outside Double Quotes (Stage Directions)
  • Physical movements and breaths: She sighs deeply., He gulps nervously., A long pause.
  • Voice adjustments: Her voice cracks., He clears his throat.
Common Pitfalls

Avoid putting explicit words like Sigh, Cough, or Gasp inside quotes. The model will literally pronounce them. Put them outside quotes as directions!