Talk naturally
Voice activity detection, push-to-talk, background-noise handling, interruptions and respectful turn-taking.
Planned integrationVOICE STUDIO · FIRST AUDIO SHOWCASE
Same words.
Different voices.
Choose a voice, press Play, and hear a real locally synthesized sample. Compare pacing, clarity, pronunciation, and personality for yourself. No account, microphone or model API is required.
Explore the voices ↓3 voices · 1 shared script · MP3 + transcript
LISTEN & COMPARE
Voice audition room
Each voice speaks the same script. These are genuine offline Piper-generated recordings; this is not a real-time conversation demo.
Nothing playing
Select a sample above.
SHARED SCRIPT
What you're hearing
The lights go dark. The city still needs you. Restore the grid, deliver the supplies, and solve the next challenge. Every choice changes the system.
THE FULL CONVERSATION EXPERIENCE
What modern voice AI should feel like
Choose your voice
Voice auditions, accents, playback controls, speaking style presets and pronunciation dictionaries.
Audition playback liveAlways show the words
Live transcripts, subtitles, captions, copyable text and clear indication of what the system understood.
Sample transcript liveSwitch conversation modes
Casual chat, tutor, game narrator, project guide, mock interviewer, co-designer and systems troubleshooter.
Persona routing plannedControl your experience
Mute, stop, replay, adjust speed, opt out of recording, and delete any stored history.
Local demo controls liveTrust what it says
Show model and voice identity, tool receipts, privacy handling, latency, uncertainty and sources.
Release standard definedNEXT IN THE PIPELINE
A studio, not just a voice dropdown
The next release should add genuine two-way conversational audio to the existing authenticated Open WebUI route, followed by verified narration exports for games and videos. A public microphone/chat endpoint must be separately rate-limited and approved before exposure.
LIVEThree recorded voice samples, transcript, sample controls, same-script comparison
NEEDS TESTSSpeech recognition, voice conversations, interruption handling, latency and turn detection
PLANNEDMore licensed voices and languages, speech-style presets, lip-sync, emotional narration, dubbing and localization
Privacy: Playing these samples does not upload microphone audio. All recordings on this page are approved static audio files. Voice cloning or personalized speech will require explicit consent and rights checks. No private administration or unrestricted AI API is exposed.