> ## Documentation Index
> Fetch the complete documentation index at: https://pathlit.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# MCP Node

> Use Model Context Protocol servers to extend AI capabilities

## Overview

The MCP (Model Context Protocol) Node enables AI agents to use tools from MCP servers. This node enables you to:

* Connect to MCP-compliant tool servers
* Use external tools and APIs through standardized interfaces
* Extend AI capabilities with specialized functions
* Integrate with services that provide MCP servers
* Access resources and tools from various providers

<Warning>
  This node is experimental and may be unreliable. Use with caution in production workflows.
</Warning>

## Configuration Parameters

### Node Configuration

* **MCP Server**:

  Select which MCP server to connect to. Available options include:

  * GitHub MCP Server (for GitHub operations)
  * Other configured MCP servers

* **Query**:

  The task or question for the agent to accomplish using the MCP server's tools

  > *What tools can you use?*

  > *Create a new issue in the repository*

  > *Search for pull requests related to bug fixes*

<AccordionGroup>
  <Accordion title="Advanced Settings">
    <ul>
      <li>
        <strong>System Prompt</strong>: Customize how the agent uses MCP tools
      </li>

      <li>
        <strong>Model</strong>: Select the language model (must support function calling)
      </li>

      <li>
        <strong>Max Steps</strong>: Maximum number of reasoning and tool-use iterations (1-50)
      </li>
    </ul>
  </Accordion>
</AccordionGroup>

## Expected Inputs and Outputs

* **Inputs**:

  * The node accepts text input that can be referenced in the query

* **Outputs**:

  * **output**: The agent's final response after using MCP tools
  * **conversation**: Detailed log of tool calls and reasoning

## Use Case Examples

1. **GitHub Automation**:

   Use the GitHub MCP server to automate repository operations like creating issues, managing pull requests, or updating code.

2. **External Service Integration**:

   Connect to MCP servers provided by third-party services for specialized functionality.

3. **Tool Discovery**:

   Query MCP servers to discover available tools and capabilities, then use them dynamically.

## Error Handling and Troubleshooting

* **MCP Server Connection**:

  Ensure the MCP server is accessible and properly configured with any required authentication.

* **Tool Compatibility**:

  Not all MCP tools may work as expected. This is an experimental feature with ongoing improvements.

* **Model Selection**:

  Use models with strong function-calling capabilities for reliable MCP tool usage.

If you encounter any issues not covered in this documentation, please reach out to our [support team](mailto:dev@pathlit.ai) for assistance.

## Relevant Nodes

<CardGroup>
  <Card title="LLM Node" href="/docs/nodes/llm">
    Standard LLM operations
  </Card>

  <Card title="GitHub Agent" href="/docs/nodes/github-agent">
    Direct GitHub integration
  </Card>

  <Card title="REST API Node" href="/docs/nodes/rest-api">
    Call REST APIs directly
  </Card>
</CardGroup>