RESEARCH
RESEARCH
Research is the whole point.
Research is the whole point.
Creatify Lab is a research group building the models behind lifelike AI video. We publish what we learn, document what our systems can and cannot do, and hold the work to one bar: realism you cannot tell from real.
Creatify Lab is a research group building the models behind lifelike AI video. We publish what we learn, document what our systems can and cannot do, and hold the work to one bar: realism you cannot tell from real.
Founded by researchers and engineers from
Founded by researchers and engineers from
META
META
SNAP
SNAP
FAIR
FAIR
ICML Test-of-Time Award
ICML Test-of-Time Award
OUR APPROACH
OUR APPROACH
Most systems animate a mouth. The harder problem is everything else: gaze, micro-expression, head motion, and holding identity and fine detail stable over time. That is where generated video still breaks, and where perceived realism is won or lost.
We treat realism as a measurable target. We evaluate models on identical inputs, track the specific failure modes that give synthesis away, and report the results, including the cases where our own models fall short.
We build models, not features. We document training data and limitations, and we would rather publish a weakness than hide it. Trust in synthetic media has to be earned in the open.
Most systems animate a mouth. The harder problem is everything else: gaze, micro-expression, head motion, and holding identity and fine detail stable over time. That is where generated video still breaks, and where perceived realism is won or lost.
We treat realism as a measurable target. We evaluate models on identical inputs, track the specific failure modes that give synthesis away, and report the results, including the cases where our own models fall short.
We build models, not features. We document training data and limitations, and we would rather publish a weakness than hide it. Trust in synthetic media has to be earned in the open.
"Our bar is simple: if a viewer can tell it is AI, the model is not finished."
THE TEAM
THE TEAM
Most systems animate a mouth. The harder problem is everything else: gaze, micro-expression, head motion, and holding identity and fine detail stable over time. That is where generated video still breaks, and where perceived realism is won or lost.
We treat realism as a measurable target. We evaluate models on identical inputs, track the specific failure modes that give synthesis away, and report the results, including the cases where our own models fall short.
We build models, not features. We document training data and limitations, and we would rather publish a weakness than hide it. Trust in synthetic media has to be earned in the open.
Most systems animate a mouth. The harder problem is everything else: gaze, micro-expression, head motion, and holding identity and fine detail stable over time. That is where generated video still breaks, and where perceived realism is won or lost.
We treat realism as a measurable target. We evaluate models on identical inputs, track the specific failure modes that give synthesis away, and report the results, including the cases where our own models fall short.
We build models, not features. We document training data and limitations, and we would rather publish a weakness than hide it. Trust in synthetic media has to be earned in the open.
CHIEF RESEARCH SCIENTIST
CHIEF RESEARCH SCIENTIST
Ledell Wu
Co-founder and Chief Research Scientist. Formerly a research scientist at Facebook AI Research (FAIR), where she led PyTorch-BigGraph and StarSpace, and a recipient of the ICML 2023 Test-of-Time Award. She leads the team building Aurora and the models that follow.
We're building the team. Wu
Researchers and engineers working on perceptual realism in generated video.
FREE REPORT
FREE REPORT
The AI Avatar Tech Report 2026.
The AI Avatar Tech Report 2026.
A deep look at how AI avatars are actually used in production: tens of thousands of generations analyzed, the buyer personas driving adoption, the use cases winning across marketing, sales, and learning, and where the technology goes next.
A deep look at how AI avatars are actually used in production: tens of thousands of generations analyzed, the buyer personas driving adoption, the use cases winning across marketing, sales, and learning, and where the technology goes next.
A deep look at how AI avatars are actually used in production: tens of thousands of generations analyzed, the buyer personas driving adoption, the use cases winning across marketing, sales, and learning, and where the technology goes next.
A deep look at how AI avatars are actually used in production: tens of thousands of generations analyzed, the buyer personas driving adoption, the use cases winning across marketing, sales, and learning, and where the technology goes next.
Drop your email and we'll send it over. No spam, just the report and the occasional research note.
The research lab building lifelike AI video models.
Creatify Lab • Copyright © 2026
The research lab building lifelike AI video models.
Creatify Lab • Copyright © 2026
The research lab building lifelike AI video models.
Creatify Lab • Copyright © 2026
The research lab building lifelike AI video models.
Creatify Lab • Copyright © 2026
