An avatar that moves when you talk.
Bring art you or your artist made, or generate a character from one prompt. They get six poses and a mouth that moves with your own voice. No webcam, no face tracking, one browser source in OBS.
Art in. Avatar out.
Bring your art
Upload the character you or your artist made, pose by pose. No character yet? Type a prompt and generate one.
Six poses, one prompt
Idle, thinking, talking, laughing, surprised, happy. Generated as one character sheet, then swapped live as they react.
Paste one URL
One browser source in OBS or Streamlabs. Transparent background, drag them onto your scene like any other layer.
Talk. The mouth follows.
No-Cam Mode syncs their mouth to your live mic. No camera pointed at you, no tracking to calibrate, and a pitch slider if you would rather not sound exactly like yourself.
Your art or ours.
The best avatar is the one that already looks like you. Upload your own art, or the piece your artist made, one image per pose. Still hunting for a design? Type a prompt and get a full character sheet back, every pose drawn to match, and make as many as you want until one feels right.
Turn on No-Cam Mode and your avatar runs on your voice alone. No camera pointed at you, no tracking software to calibrate, no lighting to fight. Bad hair day, messy room, camera shy: none of it reaches your stream. You sound like you, and your character carries the rest.
And when you want them to, they can talk back.
Sidecaster began as an AI co-host for Twitch. Flip that side on and your character reads your chat and answers out loud, hears you after a wake word, and thanks every follow, sub, and raid by name. Or leave it off and they stay purely yours: your voice, your character, nothing else in the loop.
Live in about
ten minutes.
Link your Twitch and you get 1,000 free credits, no card. Enough to make a character you actually like and have them on your scene tonight. Windows, with OBS or Streamlabs.