mcp

The Blockscout MCP Server wraps Blockscout APIs and exposes blockchain data—balances, tokens, NFTs, contract metadata—via the MCP interface so AI agents, IDEs, and automation tools can access it in a structured, context-aware way. It supports multi-chain data access by leveraging Blockscout instances and exposes a versioned REST API to interact with tools and prompts efficiently. You can use the server to query addresses, tokens, token transfers, NFT holdings, contract ABIs, and block/transaction details, all through standardized MCP tool responses. The server also includes features like intelligent context slicing, opaque cursor pagination, and automatic truncation of large fields to optimize token usage while preserving data accessibility. For practical usage, you can connect clients like Claude, Cursor, or Gemini/Gemini CLI integrations and use the provided Tool suite to inspect contracts, fetch address info, and execute read-only contract calls in a controlled way.

Prerequisites:

• Docker installed and running on your machine
• Internet access to pull the MCP server image from GitHub Container Registry

Step-by-step installation:

1. Install Docker:

# Follow OS-specific instructions at https://docs.docker.com/get-docker/

2. Pull and run the Blockscout MCP Server (latest):

docker run --rm -i ghcr.io/blockscout/mcp-server:latest

3. (Optional) Run with a named container for easier management:

docker run --name blockscout-mcp -d --restart unless-stopped -i ghcr.io/blockscout/mcp-server:latest

4. Verify the server is exposed (default port and routing may be defined by the image). If you need a specific port or environment configuration, adjust docker run options accordingly, e.g. to map port 8080:

docker run --rm -p 8080:80 ghcr.io/blockscout/mcp-server:latest

5. Connect your MCP client to the server using the provided MCP endpoint (e.g., https://mcp.blockscout.com/mcp or your local mapping) and start issuing MCP tool prompts.

If you prefer a different runtime, see the repository for any alternative installation notes or recommended patterns, but the Docker approach is the one demonstrated in the official docs.

Tips and notes:

• The server offers multi-chain data access by referencing Blockscout instances via Chainscout; ensure you configure the appropriate chain endpoints in your client if needed.
• The REST API is versioned to maintain compatibility as capabilities evolve; check API.md for exact endpoints and payload formats.
• Tool methods include address/info lookups, token/NFT queries, contract ABI retrieval, and read-only contract calls via read_contract, as detailed in the Tool Descriptions section.
• Pagination uses opaque Base64URL-encoded cursors to simplify navigation; if you encounter long results, adjust page sizes to avoid token overflows.
• Large data fields may be truncated with indicators and access guidance to fetch full data when required.
• If integrating with Claude, Cursor, or Gemini CLI, reference the respective setup sections in the README for connectors or configuration examples.
• Environment variables can be used to customize endpoints, timeouts, and other behavior; consult the server docs or SPEC.md for supported env vars and defaults.