Analysis updated 2026-05-18
Research how motion based verification could distinguish real humans from AI generated video.
Explore zero-knowledge proofs that confirm a human is present without revealing their identity.
Study a proposed protocol for AI agents to register and prove their own digital identity.
Review the technical specification and provide feedback as an early peer reviewer.
| myshapeprotocol/myshape-protocol | 0xkinno/astraea | 0xkinno/halcyon | |
|---|---|---|---|
| Stars | 0 | 0 | 0 |
| Language | TypeScript | TypeScript | TypeScript |
| Setup difficulty | — | hard | hard |
| Complexity | — | 4/5 | 4/5 |
| Audience | researcher | developer | developer |
Figures from each repo's GitHub metadata at analysis time.
MyShape is a research project trying to solve a specific problem: as AI gets better at generating fake faces, voices, and fingerprints, the usual ways of proving who someone is online are starting to break down. Instead of checking a static photo or fingerprint once, MyShape proposes checking something harder for AI to fake, the small, irregular way a real human body moves over time. The system works by pointing a camera at a person, tracking their body position with pose detection software, and turning that movement into a set of numbers that describe how the motion behaves, including its timing, acceleration, and tiny natural variations. Real human movement contains a certain amount of randomness and imperfection that the authors argue current AI cannot fully reproduce, due to limits in how cameras capture depth and motion. From that data, the system builds a small cryptographic proof, roughly 250 bytes in size, that a real human is currently present and moving, without revealing who that person actually is or sending the raw video anywhere. All of the processing described happens on the person's own device rather than on a remote server. The project calls this idea Proof of Continuity: not just proving who you are at one moment, but proving that the same sovereign digital identity has stayed unbroken over time. An early group of 100 founding participants, called the Genesis Cohort, forms what the project describes as the root trust source for the whole system, and this founding tier will not be offered again once filled. Technically, the website and supporting infrastructure are built with Next.js, React, and Tailwind CSS for the frontend, Three.js for 3D visuals, MediaPipe for tracking body pose, Supabase for the backend database, and cryptographic libraries for the zero-knowledge proofs. Several API endpoints exist for registering an identity, verifying a person's presence, and letting AI agents themselves register within the system. As of this readme, this is early stage research and the protocol's real world reliability has not yet been independently verified.
A research protocol that tries to prove a real human is present over time by analyzing the natural imperfections in their body movement, using zero-knowledge proofs.
Mainly TypeScript. The stack also includes Next.js, React, Tailwind CSS.
Mainly researcher.
This repo across BitVibe Labs
Verify against the repo before relying on details.