Blockscout MCP Server – Blockchain Data Analysis
Blockscout's official MCP server exposes blockchain explorer data through the Model Context Protocol. Use it to analyze EVM-chain addresses, transactions, tokens, NFTs, blocks, smart-contract metadata, verified source code, ABIs, and multi-chain activity from AI agents or IDEs.
Overview
Blockscout provides an official MCP server that wraps Blockscout APIs and exposes blockchain explorer data to AI agents through the Model Context Protocol. It is designed for agents, IDEs, and automation tools that need contextual access to EVM-chain data without manually parsing explorer pages or handling many raw API calls.
What the MCP server enables
The Blockscout MCP server gives AI agents blockchain-analysis tools for balances, tokens, NFTs, smart contracts, transactions, blocks, and chain metadata. The hosted server supports Streamable HTTP at the primary MCP endpoint and also exposes REST endpoints that mirror the MCP tools for systems that do not speak MCP directly.
Documented tool areas include:
- Discovering supported chains and resolving Blockscout instance URLs through Chainscout metadata.
- Resolving ENS names to Ethereum addresses.
- Looking up token symbols and token metadata.
- Retrieving smart-contract ABIs and inspecting verified contract source code.
- Getting address information, ERC-20 holdings, NFT holdings, transactions, token transfers, blocks, and related blockchain context.
- Using custom host instructions and progress notifications for longer blockchain-analysis workflows.
Blockscout also documents an optional Blockscout Analysis agent skill for more structured reasoning and analysis workflows.
When to use it
Use Blockscout MCP when an AI agent needs to analyze EVM blockchain activity. Practical examples include checking whether an address approved a token, reviewing wallet balances, inspecting contract code, resolving a smart-contract ABI for IDE work, analyzing transaction histories, checking token transfers, or exploring activity across supported chains.
It is useful for blockchain analysts, Web3 developers, security researchers, portfolio and wallet-analysis tools, and AI assistants that need structured explorer data.
Connection and authentication
The official hosted MCP endpoint is:
https://mcp.blockscout.com/mcp
Blockscout documents setup for Claude connectors, Claude Code, ChatGPT Apps, Codex, Cursor, and Gemini CLI. Example Claude Code setup uses:
claude mcp add --transport http blockscout https://mcp.blockscout.com/mcp
The hosted endpoint is documented as a public MCP server; no static API-key header is required for the basic connection. For local development, Blockscout documents running the server with Docker:
docker run --rm -i ghcr.io/blockscout/mcp-server:latest
Community telemetry can be disabled for local runs with BLOCKSCOUT_DISABLE_COMMUNITY_TELEMETRY=true.
Key considerations
The first tool call should use Blockscout's custom-instructions tool before deeper analysis, because the server documents __unlock_blockchain_analysis__ as a mandatory first step. Blockchain results may be large, so Blockscout uses pagination, response slicing, truncation indicators, and progress notifications to reduce context overflow. Always verify high-impact conclusions against source transactions or contract details before acting. Local telemetry is opt-out through an environment variable. The server is for analysis and data retrieval, not for signing transactions or taking custody of funds.
Supported Transports
streamable_http
URL: https://mcp.blockscout.com/mcp
stdio
Command: docker
Args:
run--rm-ighcr.io/blockscout/mcp-server:latest
Frequently Asked Questions
- When should an AI agent use the Blockscout MCP server?
- Use it when an agent needs structured EVM blockchain explorer data, such as checking wallet balances, analyzing transactions, inspecting token transfers, resolving ENS names, retrieving contract ABIs, reviewing verified contract code, or investigating activity across supported chains.
- What does the Blockscout MCP server add to an AI agent's capabilities?
- It gives the agent contextual access to Blockscout blockchain data through MCP tools, including balances, tokens, NFTs, contract metadata, transactions, blocks, chain lists, smart-contract ABIs, and verified contract source-code inspection.
- What can an AI agent access or manage through Blockscout MCP?
- The agent can access and analyze public blockchain explorer data such as chain metadata, addresses, ENS resolution, token information, contracts, ABIs, source code, balances, ERC-20 holdings, NFT holdings, transactions, token transfers, and blocks. It does not sign transactions or manage funds.
- How is authentication configured for the Blockscout MCP server?
- The hosted endpoint at https://mcp.blockscout.com/mcp is documented as a public MCP endpoint and does not require a static API-key header for basic connection. Local Docker development can be run without credentials, with telemetry optionally disabled by BLOCKSCOUT_DISABLE_COMMUNITY_TELEMETRY=true.
- Which transport should be used for the Blockscout MCP server?
- Use the hosted Streamable HTTP endpoint https://mcp.blockscout.com/mcp for standard client setup in Claude Code, Codex, Cursor, Gemini CLI, and other MCP clients. Use the documented Docker stdio command only for local development or self-hosted testing.