---
updatedAt: 2026-10-16T12:59:06.000Z
agentTools:
  projectIndex: https://docs.infiniaweb.com/llms.txt
---

# MCP

Connect AI code editors and assistants to the Infinia API with the Infinia MCP server.

The Infinia API Model Context Protocol (MCP) server lets AI-powered code editors such as Cursor and Windsurf, and general-purpose tools such as Claude Desktop, work directly with the Infinia API and its reference.

Model Context Protocol (MCP) is an open standard that lets AI applications securely access external data sources and tools. This page explains what the Infinia MCP server provides, how to connect your tool to it, and how to test the connection.

## What the server provides

The Infinia MCP server gives AI agents:

* **Direct API access** to Infinia API functionality.
* **API reference search** across the endpoints, parameters and data models of every Infinia API product.
* **Real-time data** from your Infinia API account, when you provide your API credentials.
* **Code generation** assistance for Infinia API integrations.

It exposes these tools to the agent:

| Tool               | What it does                                                                   |
| :----------------- | :----------------------------------------------------------------------------- |
| `list-specs`       | Lists the API reference specs, one per product.                                |
| `list-endpoints`   | Lists every path and HTTP method of a spec, with summaries.                    |
| `search-endpoints` | Searches paths, operations and parameters by keyword.                          |
| `get-endpoint`     | Returns the details of one endpoint, including its security scheme and server. |
| `execute-request`  | Sends a request to the API.                                                    |

## Set up the server

Infinia hosts a remote MCP server at `https://docs.infiniaweb.com/mcp`. Add it to your AI tool's configuration:

<Tabs>
  <Tab title="Cursor">
    **Add to `~/.cursor/mcp.json`:**

    ```json
    {
      "mcpServers": {
        "infinia-api": {
          "url": "https://docs.infiniaweb.com/mcp"
        }
      }
    }
    ```
  </Tab>

  <Tab title="Windsurf">
    **Add to `~/.codeium/windsurf/mcp_config.json`:**

    ```json
    {
      "mcpServers": {
        "infinia-api": {
          "url": "https://docs.infiniaweb.com/mcp"
        }
      }
    }
    ```
  </Tab>

  <Tab title="Claude Desktop">
    **Add to `claude_desktop_config.json`:**

    ```json
    {
      "mcpServers": {
        "infinia-api": {
          "url": "https://docs.infiniaweb.com/mcp"
        }
      }
    }
    ```
  </Tab>
</Tabs>

The Infinia API uses HTTP Basic auth. To let the agent send authenticated requests, pass your credentials as headers, either via query parameters or however your MCP client configures headers. See [Authentication](https://docs.infiniaweb.com/reference/authentication).

<Callout icon="🔒" theme="warning">
  Give your AI tool sandbox credentials only. The API reference the server reads points at the sandbox host, `app2test.infiniaweb.com`, and requests sent with production credentials move real funds.
</Callout>

## Test your setup

Once the server is configured, test the connection:

1. **Open your AI editor**, such as Cursor or Windsurf.
2. **Start a new chat** with the AI assistant.
3. **Ask about the Infinia API**, for example:
   * "How do I authenticate requests to the Infinia API?"
   * "Show me an example request to create a payout."
   * "Write a Python function that lists the movements of an account with the Infinia API."

The assistant answers from the Infinia API reference through the MCP server and, when you've provided credentials, can read your account data.