DeFi Yields MCP

What is DeFi Yields MCP?

DeFi Yields MCP is an MCP server that enables AI agents to explore DeFi yield opportunities using data from DefiLlama. It is designed for developers and users who want to integrate on-chain yield analysis into AI workflows.

How to use DeFi Yields MCP?

Install the server via uvx mcp install -m defi_yields_mcp --name "DeFi Yields Server" or manually add a configuration to Claude Desktop’s settings. Once configured, use natural language commands such as “Fetch yield pools for the Lido project” or “Analyze yield pools on Ethereum.” The server exposes the get_yield_pools tool and the analyze_yields prompt.

Key features of DeFi Yields MCP

• Retrieves DeFi yield pool data from DefiLlama.
• Filters yield pools by chain (e.g., Ethereum, Solana) or project (e.g., Lido, Aave).
• Provides an analysis prompt that guides AI agents to examine APY, 30-day mean APY, and predictions.
• Runs directly with uvx – no manual setup beyond configuration.
• Returns structured JSON output for downstream use.

Use cases of DeFi Yields MCP

• Fetch live yield pools for a specific blockchain to identify top opportunities.
• Analyze STETH pool performance including predicted APY trends.
• Compare 30‑day mean APYs across multiple Solana pools.
• Integrate DefiLlama yield data into an AI‑powered financial assistant.

FAQ from DeFi Yields MCP

What data source does DeFi Yields MCP use?

It fetches yield pool data from DefiLlama’s public API.

How do I install DeFi Yields MCP?

Run uvx mcp install -m defi_yields_mcp --name "DeFi Yields Server" for automatic installation, or manually add the server configuration to Claude Desktop’s settings file.

What tools does the server provide?

It provides the get_yield_pools tool to fetch filtered yield data and the analyze_yields prompt that generates tailored analysis instructions for AI agents.

What are the runtime requirements?

Python 3.10 or later and the uvx package runner are required.

Is authentication needed to use the server?

No authentication is required; it uses public DefiLlama data with no API keys or credentials.