CodeLLM MCPs: Unlock Superpowers with MCP Server Integration

Integrate MCP (Model Context Protocol) servers to supercharge CodeLLM with real-world context and automation. Whether you're pulling live data, triggering workflows, or connecting to your favorite tools—MCPs make your AI coding assistant infinitely more capable.

What is MCP?

Model Context Protocol (MCP) is an open standard that lets CodeLLM connect with external tools, APIs, databases, and services — all through a structured interface.

It enables the AI assistant to see, act, and reason beyond static context, allowing it to perform tasks like:

• Reading and updating live documents
• Pushing code to GitHub
• Filling out forms
• Querying databases

Think of MCPs as real-time memory extensions for your AI.

Why Use MCPs in CodeLLM?

• Real-Time Context: Feed live data from your tools directly into CodeLLM.
• Actionable Agents: Let your agent take action — push commits, run scripts, or hit APIs — all through secure tool calls.
• Custom Workflows: Tailor the AI to match your team’s exact workflow, using internal tools or 3rd-party integrations.

Where to Find MCP Servers

Explore trusted MCP server directories at: Zapier MCP and Github repos.

These directories guide you through getting the required API keys or tokens to activate each tool.

Setting Up MCP Servers in CodeLLM

1. Access MCP Configuration:

• Navigate to CodeLLM's Settings in the bottom left corner on the sidebar.
• Click on the MCP Settings under that.

• You can also search for MCPs in the top bar

  

  

2. Add a New MCP Server:

• Click Add new MCP Server

• Paste the server’s config in mcp.servers within settings.json.

  

  

3. Configure Server Details:

Configure the MCP server JSON based on its transport type:

• Stdio (Standard Input/Output): For local servers, provide the required parameters ()command, args, env, etc.)

Example:

"command": "npx",
"args": [
    "-y",
    "@modelcontextprotocol/server-github"
]

• SSE (Server-Sent Events): For remote servers, provide the server URL and other optional parameters.

Example:

http://example.com:8000/sse

<span style="display:block;text-align:center">
    ![Search for Rule](/static/imgs/docs/codellm_mcp_setup_5.webp){width=900px }
</span>

Example Config:

The example below shows the JSON configuration for two MCP servers: GitHub and Google Tasks.

• Each server config JSON starts with the server name ("github" and "google_tasks").
• Multiple server configurations are separated by a comma.

"mcp.servers":{
    "github": {
        "command": "npx",
        "args": ["-y", "@modelcontextprotocol/server-github"],
        "env": {
            "GITHUB_PERSONAL_ACCESS_TOKEN": "<YOUR_TOKEN>"
        }
    },
    "google_tasks": {
        "url": "<REMOTE_SERVER_URL>"
    }
}

4. Test the Connection:

• Go to the MCP Settings page to check the status of the servers

• Check the badge next to the server:

  • Green = Active
  • Grey = Inactive

  

Using MCPs with the Agent

Once your server is active, you can ask the agent to perform tasks like:

• “Push my latest code to GitHub”
• “Scrape this website and update my spreadsheet”
• “Trigger CI/CD pipeline with these parameters”

CodeLLM will:

• Look up the available MCP tools
• Propose a tool call
• Ask you to approve (Human-in-the-Loop)
• Execute the task via the server

You stay in control. The AI never acts without your go-ahead.

Best Practices

Security First

• Only use MCP servers from trusted sources
• Always enable authentication (OAuth, API keys, PATs)
• Follow the principle of least privilege — give only required access

Clear Server Setup

• Every MCP server must:
  • Have a unique name
  • Follow the expected config json format

FAQs and Troubleshooting

Why isn’t my MCP server connecting?

Please ensure that the command or URL in config json is correct and the json format is as shown in the example config, and verify that the server is running.

A common mistake is copying a sample config like the one below without replacing placeholder values, such as the authentication token:

"github": {
    "command": "npx",
    "args": ["-y", "@modelcontextprotocol/server-github"],
    "env": {
        "GITHUB_PERSONAL_ACCESS_TOKEN": "<YOUR_TOKEN>"
    }
}

If required authentication parameters (e.g., access tokens) are missing or not correctly filled in, the MCP server will fail to connect. Always make sure to generate and insert valid credentials as per the instructions provided on the respective MCP directory or website.

Some platforms may wrap the json config within this given structure. When copying the config, make sure to include only the server configuration JSON (i.e, <your server config json> part in below example) —not the wrapper. You may refer to the example config json above.

{
    "mcpServers": {
        <your server config json>
    }
}

Can I use multiple MCP servers?

Yes, CodeLLM supports adding multiple MCP servers, and there is no hard limit on the number of tools you can connect across them. However, the effective limit may depend on the language model (LLM) being used by the agent.

As a best practice, we recommend limiting the number of tools to ensure better performance and avoid overwhelming the LLM with too much context.

How do I secure my MCP connections?

Use MCP servers from trusted sources only and try to follow OAuth authentication in remote servers.

Need More Help?

For further assistance, please contact our support team: support@abacus.ai.