Available now
- The recording studio, running entirely in your browser
- Recorded avatar answers you can judge for yourself
- A personal walkthrough of the pilot
Platform
IRIS AVA handles the whole path: capturing good source footage, building the avatar, connecting your knowledge and showing it wherever your visitors are.
The path
You deal with one party from first recording to the avatar in your lobby. Here is what happens at each stage and where we stand today.
A browser studio that guides one of your people through a good recording: framing, light, and the listen, speak, listen rhythm the avatar is built from. Recordings stay on the device until you choose to send them.
Our renderer learns the face from your recording and generates the whole face while it speaks, eyes and teeth included. Body, hair and hands come from the real footage, with an idle loop for the moments in between. A person on our team reviews every avatar before you see it.
We connect the content you approve: product information, opening hours, directions, the questions your staff answer every day. The avatar answers from that content and declines what it does not know.
The same avatar runs life size in a Holobox, floating on a Holofan, on portrait displays and in a browser. Every screen connects through IRIS OS, so you switch it live between your content, a real colleague and your avatar, and manage the whole fleet from one place.
Run it in our cloud, on a dedicated GPU, or on premises on one NVIDIA GPU of about 32 GB. Connect your own language model through an OpenAI compatible endpoint, so conversations never leave your network.
Quality first
Most avatar tools are designed for a small video window. Ours has to hold up in The 86, where a person stands at full height in front of your visitor. That standard shapes every choice in the renderer.
Honest status
We would rather show you less and have all of it be real.
Record yourself in the studio, then tell us where your avatar should live.