Skip to main content

​
Overview

Pond3r provides a Model Context Protocol (MCP) server that lets AI agents query curated crypto datasets directly. The server uses Streamable HTTP transport, so it works with Claude Desktop, Claude Code, Cursor, and any MCP-compatible client.

​
Getting Your API Key

1

Create an account

Sign up at makeit.pond3r.xyz if you haven’t already.
2

Generate an API key

Go to makeit.pond3r.xyz/api-keys and generate a new key.
Keep your API key secure. Never expose it in client-side code or public repositories.

​
Setup with Claude Code

Run this command in your terminal to add the Pond3r MCP server: 
claude mcp add pond3r-data \
  --transport http \
  https://mcp.pond3r.xyz/mcp \
  --header "Authorization: Bearer your_api_key_here"
The tools will be available immediately in your Claude Code session.

​
Setup with Claude Desktop

Add the following to your Claude Desktop MCP configuration file (claude_desktop_config.json): 
{
  "mcpServers": {
    "pond3r": {
      "type": "http",
      "url": "https://mcp.pond3r.xyz/mcp",
      "headers": {
        "Authorization": "Bearer your_api_key_here"
      }
    }
  }
}
Restart Claude Desktop and the Pond3r tools will be available in your conversations.

​
Setup with Cursor

In Cursor, go to Settings > MCP Servers and add a new server:
  • URL: https://mcp.pond3r.xyz/mcp
  • Headers: Authorization: Bearer your_api_key_here

​
Setup with Other Clients

For any MCP-compatible client that supports Streamable HTTP transport, send POST requests to: 
https://mcp.pond3r.xyz/mcp
Pass your API key as a Bearer token in the Authorization header: 
Authorization: Bearer your_api_key_here

​
Available Tools

Once connected, your AI agent has access to three tools:
ToolDescription
list_datasetsList all available datasets and their tables
get_schemaGet column names, types, and descriptions for a dataset
queryExecute a read-only SQL query against a dataset
The agent will automatically discover schemas and available data — you just need to ask your question in natural language.

​
Query Tool

The query tool executes read-only SQL against Pond3r datasets. A few things to note:
  • SELECT only — write operations are not allowed.
  • Use bare table names — e.g. SELECT * FROM stablecoin_yields, not fully qualified paths.
  • Results are capped at 10,000 rows — use LIMIT or WHERE filters for large datasets.
  • Cost is estimated before execution — queries that exceed your tier limits are rejected before running.

​
Example Conversation

Once the MCP server is connected, you can ask your AI agent questions like: 
What are the top 5 stablecoin yields on Ethereum right now?

Show me Polymarket trades from the last 24 hours with the highest volume.

Compare USDC yields across Aave and Compound on Arbitrum.
The agent will use the list_datasets and get_schema tools to understand what’s available, then write and execute SQL via the query tool.

​
Troubleshooting

Make sure your API key is valid and correctly set. Check that your client is sending the Authorization header with every request.
The server requires bare table names (e.g. stablecoin_yields). Do not include project or dataset prefixes in your SQL.
You’ve hit your daily or monthly spend cap. Wait for the limit to reset or upgrade your plan for higher limits.