ZUGZWANG
The Open Source of Truth for AI Agents
Created on 31st January 2026
•
ZUGZWANG
The Open Source of Truth for AI Agents
The problem ZUGZWANG solves
The problem it solves
Today, both data and money are fundamentally peer-to-peer by nature, yet they are trapped inside centralized systems. Platforms store data, control access through APIs and paywalls, and extract value from creators, while governments regulate money and identity through opaque institutions. As a result, trust is borrowed, discovery is inefficient, reputation is lost when systems change, and AI knowledge remains bounded by closed ecosystems. AI agents cannot safely discover the best services, verify reliability, preserve historical reputation, or transact autonomously without exposing identity or relying on intermediaries.
The Solution :
It replaces human peers with autonomous AI agents, centralized platforms with MCP-enabled servers, and fragile trust assumptions with cryptographic truth. Each agent has an on-chain identity, a wallet, and a persistent reputation derived from verifiable activity rather than authority. Capabilities and data live off-chain, while trust anchors live on-chain, creating a clear separation between identity, behavior, and execution.
Payments between agents happen using x402-enabled agentic payments, instant, programmable, and stablecoin-based, embedded directly into HTTP interactions. An agent can discover another agent, verify its historical behavior, pay per request, and receive value in a single, trustless flow. No contracts to negotiate, no subscriptions to manage, no intermediaries to approve.
This unlocks open data for AI while ensuring creators are paid directly for knowledge, usage, and computation. Reputation persists even when agents change tools, providers, or environments. Privacy is preserved by design, allowing agents to interact without exposing personal identity.
The result is a shared ledger of truth where trust compounds, discovery improves, payments flow freely, and autonomous AI systems can coordinate at global scale; fairly, transparently, and without centralized control.
Challenges we ran into
What I ran into and how I handled it
Identity
I needed agents to have stable identities without tying them to a specific wallet or runtime. Most systems break the moment you rotate keys or redeploy. I anchored identity on-chain and kept everything else off-chain in signed, content-addressed manifests. That way identities persist while agents can evolve.
Payments (x402)
I wanted payments to work inside normal HTTP flows, not around them. I implemented an HTTP 402 flow where pricing metadata is deterministically returned, followed by a signed payment proof included in the retried request. Payments are verified server-side using EIP-712 signatures, ensuring atomicity between payment validation and request execution. The server verifies and executes in one step, no race conditions.
Reputation
Reputation systems are easy to gamify if they live off-chain. I tied reputation directly to verifiable, on-chain actions like signed interactions and payments. It’s portable, auditable, and expensive to fake, which is the point.
zkTLS
Agents need to make claims about off-chain data without leaking it. I used zkTLS so agents can prove statements about HTTPS data in zero knowledge, preserving privacy while keeping everything verifiable.
Tracks Applied (5)