The problem ZoraAgent solves

ZoraAgent addresses two important issues:

• Using Zora as a creator can be challenging. Minting, creating contracts, exploration requires smart, responsive guidance to reduce entry barrier.
• Semantically exploring NFTs on blockchain is extremely hard, pushing people towards price-oriented exploration while creativity is about context and transmission.

ZoraAgent solves these issues by providing an interface to an extremely smart and knowledgable agent that knows its way through Zora, Blockchain and it's NFTs. Users can engage in conversations with ZoraAgent through the chrome extension and ask anything. ZoraAgent will:

• Show how to use Zora
• Show how to use the SDK
• Explain concepts
• Semantically search for NFTs

Challenges we ran into

Creating an AI agent knowledgeable about blockchain, especially Zora, is challenging. We used a method called RAG to teach the AI about Zora. The AI read Zora’s documentation, we created a guide, and provided a vector database for the AI to search for relevant information based on user queries.

The second challenge was finding a decentralized vector database to ensure ZeroAgent is accessible to everyone. We used Dria’s vector databases, which store data on the blockchain and perform semantic searches to overcome this.

The third challenge was helping users explore NFTs. Zora has a keyword search, but creative searches require nuance, complexity, and context. We indexed top Zora collections, used an AI vision model to describe their cultural impact, design, and ideas, and built another vector database for semantic searches using Dria’s decentralized vector databases.