Syntra

Syntra is an intelligent blockchain data querying MCP layer. The core of the idea lies in Hypersync’s high-performance blockchain indexing that helps developers and analysts access on-chain information instantly and in human-readable way.

Syntra easily abstracts away the complexity of RPCs, manual scripts, and data wrangling. Users can request any information from tracking ERC20 transfers to decoding Uniswap swap events, everything in natural language. The chat interface provides an easy way for users to ask their blockchain-related questions in plain English, for example “Show me all USDT transfers from the last 100 blocks”. Syntra is capable of reasoning and understanding what user wants. It translates the queries into optimized blockchain data calls and brings to you, your required data.

Syntra enables next-level on-chain analytics, research, and automation.

Syntra is built using Hypersync, Next.js, FastAPI and MCP(Model Context Protocol) to provide fast, real-time blockchain queries in natural language.

• Backend: Python + FastAPI for asynchronous REST endpoints and MCP integration. Node.js and npm for frontend dependencies.
• Frontend: Next.js + React + TypeScript for a modern, responsive UI with chat interface. Tailwind CSS for rapid, responsive design.
• Blockchain Access: Hypersync API for Ethereum data indexing (transactions, blocks, ERC20 transfers, contract logs, Uniswap swaps, wallet activity).

Users type blockchain queries in the chat interface. The frontend sends requests to the backend MCP endpoint, which maps queries to the correct Python function. Functions in syntra-api/functions/ call Hypersync API to fetch data. Backend processes and formats data, applies filtering/aggregation if needed, and returns results to the frontend. React components display results in real-time with live updates and optional QuickTasks templates.

There are separate files for each query type (txn_by_id, erc20_transfers, wallet_activity, contract_logs, uniswap_swaps, etc.). Singleton in config/hypersync_client.py manages authentication and API requests efficiently. All the query functions are exposed via REST and MCP for AI-driven calls. It works on incremental fetch based on last queried block instead of constant polling. Server-side filtering and aggregation are used to prevent huge payloads from overwhelming the frontend.

Modular React components are present in components/mcp-chat/ that handle messages, input, and QuickTasks. React Context tracks theme, live updates, and query states. Predefined query templates are also given for common blockchain operations.