AI Educational Avatars
Learn from the
greatest minds.
Every face here is a real-time 3D LAM avatar — baked from a single portrait and lip-syncing live in any language, all rendering at once on the same engine that powers our live classroom.
What you're looking at
Not a video. Not a chatbot.
A teacher who talks back.
Each avatar is a real-time 3D Gaussian-splat head that listens, thinks, and speaks — turning the greatest minds in history into mentors your students can actually have a conversation with.
The faculty
Meet the mentors.
A growing roster of educational avatars, each ready to teach their own field — live.
Albert Einstein
Physics
Marie Curie
Chemistry
Ada Lovelace
Computing
Charles Darwin
Biology
Alan Turing
Computer Science
Richard Feynman
Physics
Isaac Newton
Mathematics
Nikola Tesla
Engineering
Stephen Hawking
Cosmology
Thomas Edison
Invention
Steve Jobs
Design & Product
Elon Musk
Engineering
How it works
One portrait to a living teacher.
01
Start from one portrait
A single 2.5D portrait is baked into a photoreal 3D Gaussian-splat head — no rig, no mocap, no studio.
02
Give it a voice
Any TTS or live mic stream drives ARKit blendshapes in real time, so the avatar lip-syncs every word it speaks.
03
Teach, in the browser
The avatar runs entirely in WebGL — drop it into a lesson, a tutor, or a live classroom with sub-second latency.
Real-time
Sub-second, in the browser
Gaussian-splat rendering plus streamed blendshapes mean the avatar reacts as it talks — no pre-rendered video, no server round-trips for every frame.
Any language
Speaks your classroom
Lip-sync is driven by audio, not a script — so the same avatar teaches just as fluently in Greek, English, or any language your TTS supports.
One photo in
From a portrait to a presence
Einstein, Curie, Turing, Socrates — every mentor here was baked from a single still image with the LAM 3DGS pipeline.
Have a real conversation
with a legend.
Pick a mentor, ask anything, and watch them answer you in real time.