When we started @TaironAI, we wanted to build a proof of concept for connecting a bunch of different MCP servers with different functionalities in one place to understand how Model Context Protocol could link the parts of crypto infrastructure. So we built a directory of curated open source MCP servers for chains and protocols, a playground to test things. Each one was different: some for DeFi, some for chains, some for social data. They weren’t supposed to be connected, it was just a way to prototype and see what’s possible. At some point we realized that this playground was exactly what the market needed: a single "brain" that connects major chains and protocols and makes their onchain data and functionality accessible for AI. We needed a structured way to connect data, actions, and context across protocols. So we started writing our own MCP servers from scratch. The goal was to standardize the codebase, one format, one style, one set of interfaces for every type of integration. Once that was done, we moved to a more systematic structure with three clear layers: 1️⃣ Chains: Ethereum, Base, Arbitrum, Sonic, BNB Chain, and a few smaller/newer and upcoming chains such as Plasma, HyperEVM, Telos, Sei, Monad + some non-EVMs like Phala, Algorand, Secret… 2️⃣ Liquidity protocols: Uniswap, Aerodrome, Aave, PancakeSwap, bridges, etc. 3️⃣ Analytics and data: DexScreener, CoinGecko, DeFiLlama, The Graph, explorers, and other sources. That’s the structure behind what we call the MCP Supergraph, a connected network of servers where data, liquidity, and computation meet. The Supergraph isn’t just about connecting chains, it’s about creating one layer where - agents or devs can query data, - execute onchain actions, - and build automations across protocols without rewriting integrations every time. Each MCP server is a puzzle piece with a consistent format, open interface, and verifiable output. The more pieces we add, the more powerful the network becomes. We’re focused on expanding EVM coverage first while keeping other L1s moving in parallel. Next step: bring it all together into one Supergraph endpoint where users, devs, and enterprises can interact with all the integrated chains, onchain protocols, robotics, and IoT projects through AI. Piece by piece, it’s coming together.