InspectiFI

Problem Statement

InspectiFI is a conversational blockchain analytics assistant.It lets anyone—regardless of technical depth—ask natural-language questions about addresses, transactions, or price feeds and receiveconcise, human-readable answersbacked by live on-chain data.Thanks to an MCP (Model Context Protocol) server layer and Blockscout + 1Inch we connect LLM prompts to real-time data APIs, eliminating the lag and uncertainty of stale snapshots.We can solve typical question even tricky ones, get surprised with the result!“What tokens were involved in Tx hash0x…?” “Make me a full report about this address0x…?” “What tokens were involved in hash0x…?” “Give me the current ETH, LINK, WTBC prices”Extra Fun Feature: We use Vlayer to proof if the address that you are researching is Whale or not, If you find one, you get a nice Badge!!!

Solution

InspectiFI combines real-time blockchain data with AI to deliver human-readable insights. At its core, we use Model Context Protocol (MCP) and n8n to orchestrate AI agents capable of calling live APIs like Blockscout and 1inch. These agents interpret user questions, trigger the right data source, and return clean, contextual answers.We index Ethereum Mainnet data using the Blockscout MCP, fetch price feeds via 1inch MCP. We have used Vlayer (time travel) to prove conditionally that a wallet is Whale or active. Our backend is powered by MongoDB for session and chat history, while the frontend is built with React, shadcn/ui, and Tailwind for a smooth UX.We have used n8n to route AI calls dynamically based on query type—this allows complex, multi-source answers without hardcoding logic.