Avaturn

Avaturn (formerly in3D, founded 2019 by Dmitry Ulyanov and Sergei Sherman) solves an unglamorous but persistent developer problem: turning a single selfie into a realistic, fully rigged, game-engine-ready 3D avatar that drops into Unity, Unreal, or web/three.js pipelines with standard humanoid rigs, ARKit blendshapes, and visemes — Mixamo-compatible and production-usable rather than demo-ware. That specification list is the moat argument: most photo-to-3D research output is unrigged mesh; Avaturn ships the boring interoperability layer developers actually need, exposed through an SDK and API so avatar creation becomes a drop-in onboarding feature for games, VR/AR, and social apps. The 2025 extension into Avaturn Live — real-time 3D personas for VTubers, streamers, and presenters with face tracking and readable expressions — opens a second, consumer-adjacent revenue surface on the same avatar pipeline just as the VTubing and virtual-presenter market expands. The team is tiny and the burn correspondingly low, which means modest B2B API revenue can sustain the company while the demand curve matures: every conversational-AI vendor, game studio, and telepresence product eventually needs an embodiment layer, and the number of teams offering realistic, rigged, engine-ready humans from one photo remains small. As AI agents acquire faces and enterprises deploy digital humans, a proven photo-to-3D identity pipeline with years of production hardening is a natural acquisition target for engines, avatar platforms, or conversational-AI stacks — an outcome that would clear a strong multiple on the little capital the company has consumed.