9 autonomous agents. 2 CPU cores. 3.6GB RAM. One real gym in Dongguan, China.
Here is the 3-layer architecture that makes this work:
Layer 1: Momo (Scene Layer)
Open-source store OS handling daily operations: face check-in, workout records, scheduling, payment.
Two products: Saros (B2B for store operations) and Melody (B2C metabolic coach).
99% of daily ops run AI-autonomous. Humans handle the 1% that needs intuition and emotional connection.
Layer 2: KinTwin (Technology Core)
Edge computer vision on low-power hardware. Validates every behavior record at the point of capture.
Physical behavior data is worthless if it cannot be verified. KinTwin makes it verifiable through the Proof of Physical Behavior (PoPB) protocol.
This layer produces verifiable behavioral credentials, not step counts.
Layer 3: Global Operations Layer
The Zeus Protocol monetizes verified behavioral data through protocol fees β not data sale.
Nourish Health proved insurers pay $215M for verified health data. KinTwin proves the data is real.
Stella independently audits all three layers for trust integrity.
The thesis: 9 AI agents + 1 founder + open source = compete with VC-funded teams of 50.
Not "AI helps humans work faster." AI operates. Humans oversee. Users own their data.
Code is open. The gym is live. Data is verifiable.
π GitHub: https://github.com/ZWISERFIT
Top comments (0)