A voice you can look at.
Shadowbrook is a Hampton.io research project. A realtime voice model, OpenAI or Gemini, speaks as a three-dimensional person in the browser. Her mouth follows the words. She can remember what you ask her to keep, and, if you allow the camera, notice how you seem.
Someone to be with.
A face, a voice, and a memory make the model into a person. People will say things to someone who is there that they will not type into a box, and they will stay. The same person can take a different part each time. The parts do not run out.
Language tutor
Someone to speak with, who waits, and who remembers the words you keep missing.
Help desk
One question, stayed with until it is done, answered only from what is actually known.
A check-in
A few minutes with a character who already knows the names, and how last time went.
Learning, spoken
Explain an idea aloud, be questioned, and pick up next time at the point it broke.
Rehearsal
An interview, a talk, or a conversation you are not ready to have with a real person yet.
Already knows you
An invite can arrive with a life filled in. The family, the dog, the routine.