How to Connect MCP Servers to Continue.dev

What is Continue.dev?

Continue is an open-source AI code assistant that runs in VS Code and JetBrains. It supports multiple LLM providers (OpenAI, Anthropic, Ollama, local models) and can be extended with custom context providers and tools. Since v0.9, Continue supports MCP servers natively.

Prerequisites

• Continue.dev extension installed (VS Code or JetBrains)
• An MCP server endpoint (this guide uses DataFaucet-hosted servers, but any SSE-compatible server works)
• Continue version 0.9 or later

Step 1: Locate Your Config File

Continue stores its configuration in:

• VS Code: ~/.continue/config.yaml (or config.json for older versions)
• JetBrains: ~/.continue/config.yaml

If you are using the newer YAML format (recommended), open config.yaml. If you have config.json, the structure is similar but in JSON.

Step 2: Add Your MCP Server

Add an mcpServers section to your config:

mcpServers:
  - name: my-crm
    url: https://your-server.datafaucet.dev/sse
    headers:
      Authorization: Bearer your-api-key

For JSON config:

{
  "mcpServers": [
    {
      "name": "my-crm",
      "url": "https://your-server.datafaucet.dev/sse",
      "headers": {
        "Authorization": "Bearer your-api-key"
      }
    }
  ]
}

Step 3: Reload Continue

After saving the config, reload the Continue extension:

• VS Code: Open command palette (Ctrl+Shift+P) → "Continue: Reload Configuration"
• JetBrains: Restart the IDE or use the Continue sidebar refresh button

Step 4: Verify the Connection

Open the Continue chat panel and ask it to use one of your MCP tools:

> "Use my-crm to get the latest deals"

If configured correctly, Continue will call the MCP server and return results inline. You should see the tool call in the chat history.

Adding Multiple Servers

Continue supports multiple MCP servers simultaneously:

mcpServers:
  - name: crm
    url: https://crm-server.datafaucet.dev/sse
    headers:
      Authorization: Bearer key-1
  - name: monitoring
    url: https://monitoring-server.datafaucet.dev/sse
    headers:
      Authorization: Bearer key-2
  - name: tickets
    url: https://tickets-server.datafaucet.dev/sse
    headers:
      Authorization: Bearer key-3

The AI assistant can call any of these servers based on the question you ask.

Troubleshooting

"Tool not found" error: Make sure the server name in your config matches what you reference in conversation. Reload config after changes.

Connection timeout: Verify the server URL is correct and accessible. Test by opening the SSE URL in a browser (you should see an event stream).

Auth errors: Double-check your API key. DataFaucet keys are shown on the dashboard after deployment.

Why Continue + MCP?

Continue is open-source and model-agnostic. Adding MCP servers means you get tool access regardless of which LLM provider you use. Switch between OpenAI, Anthropic, or a local Ollama model and your tool connections persist.

Create your MCP servers with DataFaucet and connect them to Continue.dev in under 2 minutes.

Related: Connect MCP Servers to Cline, Connect MCP Servers to Roo Code, Best MCP Clients in 2026.