When AI agents query human behavioral data, the humans whose data created value should get paid. We are building the standard that makes that provable and automatic.
This document describes what we are building, what we have proven so far, and what we need from a chain partner to bring it to production.
We have distribution and a working data pipeline. 7,500 people signed up through a referral and leaderboard mechanism. 20 beta users connected their Spotify accounts. We captured 108,000 listening minutes of real behavioral data. The production product is now in development.
Music listening is the first data vertical — frequent, authentic, and easy to understand. The beta proved the data capture pipeline works. The production product will add behavioral fingerprinting, taste matchmaking, and the compensation layer on top.
An AI agent queries across a graph of human behavioral data. A zero-knowledge proof is generated that attests: these specific wallet addresses were touched in this query. The proof gets verified on-chain. Settlement reads the verified array and distributes fractionalized compensation automatically.
The behavioral data never touches the chain. The graph structure is never revealed. The humans don't have to trust the platform. The agents don't have to trust the platform. The proof does the work.
Graph traversal happened — without revealing the graph. A set of wallet addresses contributed to a query response, and a specific quantity of data entries were consumed. Nothing else is disclosed.
This is not consent management. This is not a data marketplace. This is a cryptographically enforced compensation rail for human data queries. Agents pay. Proof enforces who gets paid.
Everything above Layer 2 is conventional software. The blockchain enforces exactly one thing: that compensation is distributed to the proven wallet array.
We are evaluating chain partners based on technical fit, not ecosystem loyalty. The chain that meets the most requirements earns first-implementation credit on the standard.
A proving backend that can handle a custom ZK circuit at query scale with acceptable latency and cost.
Fractionalized micropayments to many wallets per query. Settlement cost must be negligible relative to payment value.
Behavioral data must remain private during query execution. The chain partner should support or enable privacy-preserving computation.
Agents need to call the payment function directly. A native or well-integrated machine-to-machine payment protocol is a strong advantage.
Capital to fund the pilot — circuit specification, proof generation integration, testnet deployment, and a live compensation loop with real users. Milestone-gated. Deliverables are open-source.
This is an ERC-level standard for machine-to-human compensation. The first chain to deploy it owns a permanent position in that standard's history.
We are not chain-exclusive. The standard will be multi-chain. What we are offering is first-implementation — the reference, the co-authorship credit, and the consumer proof point. That position is available once.
We are early. The data pipeline is proven. The primitive is defined. The production product, circuit specification, and chain integration are the next milestones. We are looking for a chain partner who sees the value of this standard and wants to build alongside us.
This is a conversation about technical fit and mutual value. If the fit is right, we move fast.
Machines are learning to pay each other. The question is whether the humans on the other side of the data get paid too. We think the answer should be provable.