About Us
It's not a bug, it's a feature.
This system was born from a crossover between Hollywood film sets and 3D modeling logic.
Our founding team has years of experience in film and television makeup and styling, alongside a background in 3D modeling. These two seemingly unrelated fields share a surprisingly identical underlying logic. In 3D modeling, we rarely alter the underlying base mesh to change a character's look; instead, we apply "normal maps" to alter how light reflects off the surface, instantly reshaping the visual texture.
Makeup works exactly the same way. An actor on set cannot get surgery (altering the base mesh) for every role. Find the right direction—change the brow shape, the hairline, and the color palette (changing the maps)—and the same person reads completely differently. All of it can be done in minutes.
However, we noticed an absurd disconnect.
On set, professionals always look at a face first, then decide what to do. But off set, almost every beauty tutorial online works the other way around—first it tells you what is trending, then it tells you to squeeze your face into it.
Under this "standard beauty," any trait that doesn't fit the trending template is treated as a bug to be fixed. But no one should treat their own face as a mistake.
In real professional styling, your freckles, texture, and unique bone structure are the exact features that give you absolute recognizability. What makes people remember you is never how close you are to some supposed perfect answer, but these features that belong only to you.
Over the past 8 years, our team has taught makeup offline to over 1,000 students, hoping to help more people find their right direction. But offline coaching is strictly limited by time and energy, meaning only a limited number of people can be served.
Now, with the help of AI computer vision, technology can finally scale this expertise. Systematizing this experience—which previously could only be taught one-on-one—and breaking the limits of time to serve everyone is our core mission.
We poured all the teaching experience, judgment logic, and facial data accumulated over these 8 years and our time in film into this system. That is why our algorithm can break down your facial structure like a real stylist would. We will not give you an "attractiveness score," and we will not rank you against a crowd average, because your eye spacing, face shape, and contrast are not a list of defects.
We do exactly one thing: measure your face first, get real structural data, then tell you which direction to draw your brows, what saturation of color to choose, and how to shape your hairline.
We do not stick empty labels like "soft" or "cool" on you. We only give concrete, data-backed actions.
Current AI models: we use xAI Grok 4.5 to generate written reports and ByteDance Seedream 5.0 Pro to generate model portraits and occasion styling images. We route these requests through OpenRouter. We may change models as the Service develops and will update this page when the current models change.
Because every face deserves its own manual.