Hive Intelligence home
Hive Intelligence
GitHub

Compare DeFi yields and protocol health from Cursor

Cursor speaks MCP natively as of v0.42. Once Hive is installed via mcp.json, Cursor can call get_yield_pools and get_protocol_tvl across DeFiLlama-backed data in a single conversation — returning APY, TVL, chain, asset composition, and a derived APY×TVL score for sortability. This lets a developer prototyping a yield agent ask 'find me 5 USDC pools with TVL over $100M and APY over 6%' and get a ranked answer inline. CoinGecko MCP and Moralis Cortex don't aggregate protocol-wide DeFi yields; agents using those would have to call DeFiLlama's REST surface separately and reconcile schemas.

Client: Cursor · Workflow: DeFi yield research

Hive tools used

  • get_yield_pools — Returns DeFiLlama yield-pool data with APY, TVL, chain, asset composition, and pool stability indicators.
  • get_protocol_tvl — Protocol-wide TVL with chain breakdowns, time-series, and category classification (lending, AMM, restaking, etc.).
  • get_yield_pool_chart — Yield-pool time series for checking whether APY and TVL are stable enough to trust.

Steps

  1. Add Hive to your Cursor mcp.json

    Open Cursor Settings -> MCP Servers -> Add Server, or edit ~/.cursor/mcp.json directly. Use the remote HTTP config from the Cursor install guide so Cursor connects to Hive without spawning a local Node wrapper.

    {
  "mcpServers": {
    "hive": {
      "url": "https://mcp.hiveintelligence.xyz/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_HIVE_API_KEY"
      }
    }
  }
}
  2. Confirm the tool surface in chat

    Open Cursor chat (Cmd-L), type "list available Hive tools". Cursor should see Hive's compact discovery tools; use search_tools or hive://tools for the wider catalog. If Hive is missing entirely, re-check the mcp.json URL and Authorization header.

  3. Ask Cursor to rank yield pools

    Cursor formulates the right tool calls automatically. For example, "find USDC yield pools with TVL over $100M and APY over 6%" makes it call get_yield_pools, filter, sort, and return a structured table.

  4. Drill into a specific protocol

    Once a pool catches your eye, ask "what's the protocol TVL trend for {protocol} over the last 90 days?" — Cursor calls get_protocol_tvl and reports the trajectory, flagging concentration risk if dominant chain TVL is over 70%.

  5. Generate yield-agent code from the conversation

    Because you're in Cursor, you can ask it to "write a TypeScript function that calls Hive get_yield_pools, filters by APY > 6 and TVL > 100M, and returns the top 10 sorted by APY×TVL." Cursor writes the function with the correct Hive REST signature and example output.

Example prompt

Find me 5 USDC yield pools across all chains with TVL over $100M and APY over 6%, ranked by APY×TVL. Include chain and protocol name.

Example output shape

Example output shape (illustrative values; run the tools for live data):

I queried Hive's yield-pool tools and filtered for USDC pools with TVL over $100M and APY over 6%.

| Rank | Protocol | Chain | APY | TVL | APY x TVL |
|------|----------|-------|-----|-----|-----------|
| 1 | Aave v3 USDC | Ethereum | 6.4% | $1.2B | 76.8M |
| 2 | Compound v3 USDC | Base | 7.1% | $480M | 34.1M |
| 3 | Morpho USDC | Ethereum | 6.8% | $310M | 21.1M |
| 4 | Aave v3 USDC | Polygon | 6.2% | $230M | 14.3M |
| 5 | Spark USDC | Ethereum | 7.4% | $180M | 13.3M |

Before using a pool, inspect the live pool chart, protocol TVL trend, asset composition, lockups, and smart-contract risk.

Related glossary terms

Related recipes

See also