How to Connect MCP Servers to Amazon Q Developer

Amazon Q Developer Supports MCP

Amazon Q Developer is AWS's AI coding assistant available in VS Code and JetBrains IDEs. It supports MCP servers, letting Q call external tools during coding tasks. When Q needs data from a monitoring dashboard, CRM, or internal API, it invokes MCP tools without you switching context.

Prerequisites

• VS Code or JetBrains IDE with Amazon Q Developer extension
• An MCP server URL (SSE endpoint)
• Active AWS Builder ID or IAM Identity Center credentials

Step 1: Open Amazon Q MCP Configuration

In VS Code:

1. Open Command Palette (Cmd/Ctrl + Shift + P)
2. Type "Amazon Q: Open MCP Configuration"
3. This opens your MCP settings file

Or edit the configuration file directly at ~/.aws/amazonq/mcp.json:

{
  "mcpServers": {
    "my-server": {
      "type": "sse",
      "url": "https://your-mcp-endpoint.datafaucet.dev/sse",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY"
      }
    }
  }
}

Step 2: Verify Connection

After saving the configuration, Amazon Q discovers the available tools. You can verify by asking:

> "@workspace What MCP tools are available?"

Q should list all tools from your connected servers with their names and descriptions.

Step 3: Use Tools in Your Workflow

Amazon Q calls MCP tools when relevant to your prompt. Examples:

• "Check the CloudWatch error rate for the payment service before I refactor this handler"
• "Pull the customer's account details from our admin tool"
• "Get the latest deployment status from our CI/CD pipeline"

Q decides when to call tools based on context. You can also explicitly request tool usage.

Connecting Multiple Servers

Add multiple entries to the mcpServers object:

{
  "mcpServers": {
    "monitoring": {
      "type": "sse",
      "url": "https://monitoring.datafaucet.dev/sse",
      "headers": { "Authorization": "Bearer KEY_1" }
    },
    "internal-admin": {
      "type": "sse",
      "url": "https://admin.datafaucet.dev/sse",
      "headers": { "Authorization": "Bearer KEY_2" }
    },
    "jira": {
      "type": "sse",
      "url": "https://jira.datafaucet.dev/sse",
      "headers": { "Authorization": "Bearer KEY_3" }
    }
  }
}

Q can call tools from any connected server in the same session.

Creating MCP Servers for Amazon Q

Need MCP servers to connect? DataFaucet generates them from any web app:

1. Enter the app URL in DataFaucet
2. Browse for 30-60 seconds in the cloud browser
3. API calls are captured and converted to typed tool definitions
4. Deploy with one click and copy the config into Amazon Q

Works with internal AWS tools, third-party SaaS, and any web app with an API behind its UI.

AWS-Specific Use Cases

Amazon Q paired with MCP servers is powerful for AWS-heavy teams:

• CloudWatch → Q: "What's the p99 latency for the orders API in the last hour?"
• AWS Console tools → Q: Browse internal AWS dashboards, capture their APIs, give Q access
• Cost Explorer → Q: "How much did our Lambda functions cost this week vs last week?"
• Internal microservices → Q: Browse your service mesh UI, capture endpoints, let Q query service health

Tips

• Config file location: ~/.aws/amazonq/mcp.json (shared across VS Code and JetBrains)
• Restart after changes: Amazon Q may need a reload after config updates
• Keep servers focused: One server per tool/service is cleaner than one massive server
• Auth rotation: If tools stop working, check if API keys or sessions have expired