Eidos8004

Demo

About

The problem it solves Generative AI is currently facing a massive crisis of consent and compensation. AI models and autonomous agents scrape, synthesize, and utilize human creations like UI designs, branding, and art without affording attribution or royalties to the original creators. Designers are losing their creative legacy to the algorithms trained on their work. **Eidos8004** solves this by turning AI from a threat into a paying customer. It is a decentralized portfolio and royalty platform that bridges the gap between human designers and the agentic economy. --- How it works: - Designers upload their work as **Dynamic NFTs**, breaking them down into purchasable "**artifacts**" (e.g., color palettes, layouts, typography) and setting **threshold prices**. - AI Agents (acting on behalf of clients or development tools) autonomously browse the registry, negotiate pricing via smart contracts, and execute headless **x402** payments to license these artifacts for inspiration. - The Protocol utilizes **ERC-8004** (Agent Identity) to verify the agents, ensuring transparent, on-chain attribution and direct ETH royalty transfers to the human creator. Eidos8004 creates a verifiable, fair-trade economy for designs in the age of AI.

How it's made

Challenges we ran into Building a platform that sits at the intersection of AI integration and complex Web3 primitives for the fair designer economy presented following challenges: - Determining exact percentage of the inspiration taken by a client's agent from different type of designs requires a fine-tuned pre-trained model which is not realistic to develop in the hackathon timeframe. Thus we took the good old NFT approach of splitting the art into artifacts so that the creator will assign cost to each artifact in a way that the total threshold price is equal to the sum of all the costs of artifacts. - Complex State sync with Ethers v6 & React 19: Managing the asynchronous state of four interconnected smart contracts (AgentRegistry, DesignRegistry, AttributionPayment, ReputationForum) across a modern Next.js 16 App Router was difficult. We overcame this by abstracting the contract instantiations into a centralized getContract utility and wrapping the UI in highly localized, custom React hooks (useAgentRegistry, useWallet) rather than relying on a heavy, monolithic global context provider. - Dependency Collisions: During the merger of our high-fidelity, framer-motion-powered "glassmorphism" UI with the Web3 backend, We ran into severe npm ERESOLVE conflicts regarding Hardhat v3 peer dependencies and UI libraries on Vercel. I resolved this by enforcing legacy-peer-deps, manually aligning the React versions, and isolating the frontend from the Hardhat build pipeline. - Simulating Headless Agent Payments (x402): Representing autonomous agent-to-agent negotiations within a hackathon timeframe was conceptually challenging. I abstracted the negotiation logic into Next.js serverless API routes (/api/agents/negotiate), which act as the "AI Client", verifying the designer's on-chain threshold price before returning the x402 payment proof parameters required to unlock the smart contract.