Our approach

About MyTypeFYI.fun

A type name is useful only when you can see what it refers to. MyTypeFYI.fun puts the avatars and the visible features behind a result alongside the label.

A small space for visual self-reflection

This project is for adults aged 18 and over who want to explore patterns in the faces they choose. All test faces are AI-generated virtual adults. We do not analyze your face or recommend real people.

Evidence before sweeping claims

Results use your own choices and keep picked faces, predictions, contrasts, and support counts beside the label. Simulated popularity is labeled synthetic and is never presented as real-user opinion, character evidence, or relationship evidence.

An independent project

MyTypeFYI.fun is independently developed and is not affiliated with or endorsed by mytype.fyi. We use our own avatar assets, definitions, copy, and implementation.

A method that can be inspected

The test presents 12 portraits at a time and replaces each liked position with an unseen face. Across 20 base crowds, the session compares selected and unselected faces within the same crowd. Every fifth crowd is held out as a small validation check, while reviewed prototype groups connect the learned feature pattern to the 12 launch labels.

The result keeps the explanation visible: picked portraits, predicted matches, contrasting examples, recurring features, and support counts remain beside the primary or nearby label. When support is weak or spread across several directions, the interface can return a mixed result instead of forcing certainty.

A deliberately narrow scope

The complete basic test works without an account, checkout, selfie, or server-side test history. An optional one-time Deep Report purchase appears only after a result. Free progress stays in the current browser, and the user chooses whether to download or share an exported copy.

Purchased Deep Reports can be restored on another device through the checkout email. There is no public profile, real-user population comparison, friend matching, or personalized image generation. The report uses a fixed synthetic baseline only as a clearly labeled visual reference.

Definitions are reviewed with their boundaries

Each launch type includes visible traits, nearby looks, representative virtual adults, and a statement about what the label cannot establish. The portrait catalog is reviewed before it becomes test input, and prototype groups contain several distinct faces rather than one idealized example.

Review does not remove every synthetic-image artifact or make a visual label objective. It provides a clearer, inspectable basis for the result and a way to report a portrait or content concern.

Explore the types