How to Connect MCP Servers to Claude Code

Claude Code Supports MCP Natively

Claude Code (Anthropic's CLI agent) supports MCP servers out of the box. Add any MCP server endpoint and Claude Code discovers available tools automatically during conversations.

This means your terminal-based AI agent can query databases, check dashboards, create tickets, and interact with any API you expose through an MCP server.

Prerequisites

• Claude Code installed (npm install -g @anthropic-ai/claude-code)
• An MCP server endpoint (SSE or stdio)
• A DataFaucet account (free tier works) or any MCP server URL

Configuration

Claude Code reads MCP server config from .mcp.json in your project root or ~/.claude/.mcp.json globally.

Project-level config

Create .mcp.json in your repo root:

{
  "mcpServers": {
    "your-app": {
      "url": "https://datafaucet.dev/api/mcp/YOUR_SERVER_ID/sse"
    }
  }
}

Global config

For servers you want available in every project:

{
  "mcpServers": {
    "stripe": {
      "url": "https://datafaucet.dev/api/mcp/STRIPE_ID/sse"
    },
    "linear": {
      "url": "https://datafaucet.dev/api/mcp/LINEAR_ID/sse"
    }
  }
}

Testing the Connection

Start Claude Code in your project directory:

claude

Ask something that requires your connected tool:

• "What tickets are assigned to me in Linear?"
• "Show me this month's revenue from Stripe"
• "What's the status of the latest deployment?"

Claude Code discovers tools from the MCP server and calls them inline. No extra prompting needed.

SSE vs stdio

Claude Code supports both transport types:

DataFaucet servers use SSE by default. No local process needed.

Multiple Servers

Add as many servers as you need. Claude Code connects to all of them and picks the right tool based on your question:

{
  "mcpServers": {
    "stripe": { "url": "https://datafaucet.dev/api/mcp/STRIPE_ID/sse" },
    "jira": { "url": "https://datafaucet.dev/api/mcp/JIRA_ID/sse" },
    "postgres": { "command": "npx", "args": ["@modelcontextprotocol/server-postgres", "postgresql://localhost/mydb"] }
  }
}

DataFaucet + Claude Code Workflow

1. Go to DataFaucet and enter any URL
2. Browse the app for 30 seconds (API calls get captured)
3. Deploy your MCP server (one click)
4. Copy the SSE URL into your .mcp.json
5. Start Claude Code and ask questions about that app's data

Total setup: under 2 minutes from zero to Claude Code querying a new API.

Common Patterns

Morning standup prep:

"Summarize what shipped yesterday in Linear, any failed payments in Stripe, and unread messages in the support channel."

Code review context:

"Pull the Jira ticket for this PR and summarize the requirements. Does my implementation match?"

Incident response:

"Check the last 5 error logs, query the affected customer's account in Stripe, and create a Linear ticket for the fix."

Troubleshooting

Server not connecting: Verify the URL is reachable with curl https://datafaucet.dev/api/mcp/YOUR_ID/sse. You should see an SSE stream open.

Tools not discovered: Restart Claude Code after editing .mcp.json. The config is read at startup.

Auth errors: DataFaucet servers handle auth internally (captured during browse session). No API keys needed in your config.

Next Steps

• Create your first MCP server (free, 60 seconds)
• Best MCP clients compared
• MCP servers vs function calling