Model Context Protocol

Bud’s Model Context Protocol (MCP) server offers a set of endpoints that AI agents can use to interact with Bud’s APIs as well as read our API documentation and integration guides.

If you’re using AI-powered code editors such as Cursor or Windsurf, or general-purpose tools like Claude Desktop, you can connect to the MCP server.

Setup

Authentication

The Bud MCP server is designed to be used with single-customer authentication.

First, call the POST bud-api-mcp/url endpoint with your standard OAuth Token for the Bud Platform.

You will then receive a custom MCP URL and single-customer authentication token which you can use to connect to the Bud MCP server

{
  "url": "<custom-url-mcp>",
  "token": "<customer-specific-bearer-token>"
}

Connection

You can now use this URL and Bearer Token to connect to the Bud MCP server remotely. An example configuration (e.g. for a FastMCP client) would look like this:

{
  "transport": "http",
  "url": "<custom-url-mcp>",
  "headers": {
    "Authorization": "Bearer token"
  },
  "auth": "<customer-specific-bearer-token>"
}

Tools

The server provides access to the following endpoints as tools. We recommend enabling human confirmation before using them and exercising caution when integrating Bud’s MCP with other servers to reduce the risk of prompt injection attacks.

Note: The Bud MCP server is currently in beta, and this list of tools (and their structured responses) is liable to change without the standard deprecation schedule. In particular structured responses may change in future to keep semantic information with reduced output token usage, and some query parameters may not currently be available for tool calling.

If you have any questions, please contact us via the chatbot (bottom-right of screen 👉) or via a support request or check our FAQs.