The character vault stores a generated character so the same face, body and wardrobe can be pulled into any later graph as a reference, instead of regenerating and getting a slightly different person each time. One character image supports unlimited talking head clips.
Text to image gives you a different person every run. For an account that posts daily, that reads as a rotating cast of strangers, which is the fastest way to lose the trust that makes UGC work at all.
Reference-controlled generation fixes it: the saved character image is passed as a reference on every later call, so the model matches the face instead of inventing one.
AI character, AI voice, AI lipsync. No camera, no studio, no talent fees, no release forms. The same vault entry can be put in a new outfit, a new setting or a new product scene while staying recognisably the same person.
Yes. Generated characters, generated voice and lipsync produce a complete faceless UGC ad from start to finish.
No. The vault is free and storing a character costs nothing. You only pay your provider for the generations that use it.
Yes, through ElevenLabs or Minimax on your own key. Saved voices are reusable the same way characters are.
Start free, no card
Connect one API key and generate on the canvas. Pro is $9.67/mo billed every 3 months when you want the editor, templates and Discord.