EasyOrder

Demo

About

The problem it solves Our project makes for a better orderbook by taking profits made on trades with different price levels and giving them back to users in the form of a reward token, which is different to TradFi’s approach to orderbooks, which is to pocket the money. See images submitted to get a visual aid. Liquidity pools require liquidity, and those providing liquidity risk impermanent loss, but our orderbook can matchmake orders while still being decentralized and verifiable with AVS. Our project utilizes an AI agent to analyze all of the orders on the orderbook and help the user make the best possible trade. User Interaction and Data Flow Our user visits our frontend website, where they can the use the navbar to navigate to two pages, our trading page titled “Limit” and AI chat page called “Chat”. In our “Limit” trading page, the user can make a order for any amount of Token A for any amount of Token B. In our project, we added WETH to USDC and WBTC to USDC trading pairs. The user also has an AI suggestion button that will suggest prices based on outstanding and finalized orders on chain. In our “Chat” page, our user can ask our AI chatbot agent questions about the orderbook on chain, including asking for all of the orders that have occurred on chain, asking about the best price for an order, and asking about the average price based on the most recent orders. The project architecture and development process (See image diagram submitted for visual aid for entire system architecture) For the Eigenlayer AVS: Our frontend queries our EVM smart contract to place orders. On the smart contract, when a order is created, it is stored on the smart contract, and the smart contract emits an event to a Tangle-Eigenlayer AVS system. This AVS system does match making with the order and other existing orders. If the AVS finds a match, a response function is called from the AVS with the match order, and the order is finalized. If a match cannot be made, the order is stored on chain. For Eigenlayer/Order matching smart contract: Once receiving the finalized order, the smart contract will execute the swap, and any profits made from the swap, where a buyer pays more than a seller is willing to sell for, are sent to be locked in a ERC20 Rewards token. Then, each user will get 1 reward token, incentivising them to use our orderbook to make a trade. See images submitted to get a visual aid. For the AI Agent: The AI agent is an Eliza OS framework hosted on Autonome’s AI agent hosting service (https://dev.autonome.fun/autonome). The Eliza framework does not host the AI model itself, but instead queries a Gaia node hosted on one of our laptops, which runs the AI model, DeepSeekR1, to generate responses. We wrote a custom Eliza OS plugin to interact with our orderbook and retrieve data from it, including accessing all outstanding and finalized orders. The plugin then provides examples to the AI of when to execute queries to the smart contract based on whether the user is asking about prices of tokens on the orderbook. Product Integrations Product integrations: Autonome (https://dev.autonome.fun/autonome) * Hosted Eliza OS AI agent framework Gaia (https://www.gaianet.ai/) * Hosted actual AI that Eliza OS queried ElizaOS Framework (https://github.com/elizaOS/eliza) * Framework for creating actions that an AI model can execute based on user needs. We used the framework to create a custom plugin that could interact with our orderbook on chain. Tangle/Eigenlayer AVS (https://docs.tangle.tools/) * Our AVS hosted the matchmaking execution for our orderbook, making our orderbook incredibly efficient in terms of gas fees. NextJS * Used to build out our frontend into an awesome, sleek look Docker container * Used to host our Eliza OS build which was then deployed to Autonome. Key differentiators and uniqueness of the project Our project’s orderbook is unique in that, if MEV occurs in the form of a buyer paying more than they should or seller low balling themselves, the profit made from that MEV goes directly to the user in the form of a reward token. See images submitted to get a visual aid. Our project also includes an entire AI agent setup to help the user make the best trades using the orderbook. It can easily help the user understand the orders that have previously been finalized or are still outstanding, and give the user the best price to put on their order. We imagine many orderbook projects, if there were any others, did not do this. Trade-offs and shortcuts while building Tangle’s AVS framework was very difficult to use while building our AVS, in part because of the documentation. We worked very hard with Drew, the founder of Tangle, to build our AVS and make sure the documentation would get improved in the future. In order to save time, we focused on getting our AVS and smart contract to work on a local testnet, and did not have time to deploy to a remote testnet or a mainnet. Additional Features Our project was entirely new and built from scratch. We knew nothing about AI agents, Autonome, Gaia, Tangle, or Eigenlayer before coming to the hackathon. We learned a lot during our project! Built for the Eigen Games Hackathon