// Voice-to-Voice Simulation Engine — Build 026.2
Practice speaking
to characters that
talk back.
Soliloquy is a real-time roleplay cockpit. Speak into your mic, and an emotionally responsive AI entity answers in its own voice — shifting tone, strategy, and objective based on what you say. Built for voice actors, dungeon masters, and trainers who need their next reply out loud.
Recently cloned by community
What everyone's rehearsing this week.
// Built for three rooms
Whoever is on the other side
of the microphone.
Voice Actors
Cycle through brutal character transitions on demand — noir detective, alien diplomat, panicked civilian — without booking a coach.
TTRPG / NPCs
Auto-generate non-player characters with consistent voice and motivation. Run a tavern of five distinct personalities solo.
Corporate Training
Validate front-line employees against difficult customers, hostile media, or skeptical CFOs — measured, repeatable, voice-first.
// The engine
Four channels.
One conversation.
Mic capture
Push-to-talk in browser. Webm streamed to engine.
Whisper-1 STT
Your speech is transcribed in milliseconds.
Claude Sonnet 4.5
Persona-aware reply with vocal direction, in 1–3 sentences.
OpenAI TTS
Synth voice plays back — visualized in real time on the cockpit.