This week, we built a unified AI agent interface, prototyped a manager and developer (AgentKit), and started a Gateway for chat and agent interactions. Design work is underway.
Prize Pool
We want create platform that allows users to assemble their own team consisting exclusively of AI agents to implement their projects Launching Web3 projects, even the most typical ones, often takes months, with most of the work being routine. Our platform changes this process:
We developed the backend using C# and JS. We used JS for writing agents because there are more ready-made libraries available for JS. For example, we used AgentKit.js to create an agent responsible for wallet-related actions. We also used Langchain everywhere as the foundation for any agent interactions. In C#, we wrote the main middleware layer. Our plan is that the AI agent functions as a microservice, which connects to the C# server and interacts with it through a unified interface.