Develocity Analytics MCP Server User Manual

Please contact the Develocity support team to obtain the full setup instructions to get started with Develocity Analytics MCP server.

Introduction

Develocity version 2025.4 and Develocity Reporting Kit 2.1 introduce the ability to use agentic AI applications to query data exported using the Develocity Reporting and Visualization capability, via an MCP server.

This guide provides instructions for connecting applications to and using Develocity Analytics MCP server. After completing the setup, your agentic AI applications will be able to query your Develocity Reporting and Visualization data in response to your prompts.

Enabling Develocity Analytics MCP Server

To enable this feature on your Develocity or Develocity Reporting Kit instance, contact Develocity support. You will need to be running Develocity 2025.4 or higher, or Develocity Reporting Kit 2.1 or higher.

Configuring your agentic AI application

Develocity Analytics MCP server can be used with any client application which supports the streamable HTTP transport, and MCP tools and prompts. Consult the official MCP feature matrix to see if your chosen application provides this support.

Each agentic AI application has its own method of configuring MCP servers. This guide has instructions on how to configure a number of popular applications. It is possible to use the MCP server with applications that are not listed, but these have not been tested by Gradle. Contact Develocity support if you encounter any issues.

Client applications will need to configure the MCP server as a remote HTTP-based server (as opposed to a local stdio server). Use the address of your Develocity or Develocity Reporting Kit instance, with the path /drv-mcp for the URL, for example, https://develocity.example.com/drv-mcp or https://develocity-reporting-kit.examples.com/drv-mcp.

To connect your client application to the MCP server, you’ll need a Develocity access key for a user with the necessary permissions. This key must be sent with every request inside an Authorization HTTP header.

There are two required permissions: - Users must have the Access build data via the API permission. - For Develocity instances with Project level access control enabled, users must also have the Access all data with or without an associated project permission.

Requests without a valid key or where the associated user does not have the required permissions will be rejected.

You can check which permissions your user has by visiting the /settings/access page when logged in to Develocity.

To generate an access key, follow the instructions in the Develocity API user manual.

Visual Studio Code
GitHub Copilot plugin for IntelliJ IDE
Gemini CLI
Claude Desktop
Other applications

To configure the MCP server in Visual Studio Code:

Open Visual Studio Code.
Press Ctrl+Shift+P (or Cmd+Shift+P on macOS) to open the Command Palette.
Type MCP: Open User Configuration and select it.
Paste the following JSON configuration into the appropriate section, replacing <base-url> with your Develocity or Develocity Reporting Kit URL.

Using inputs in this way is recommended to avoid hardcoding the Develocity access key within the JSON configuration. You will be prompted for your access key the first time you start the server. Consult the Visual Studio Code documentation for more details.

{
  "inputs": [
    {
      "type": "promptString",
      "id": "develocity-access-key",
      "description": "Develocity Access Key",
      "password": true
    }
  ],
  "servers": {
    "drv-ai": {
      "url": "<base-url>/drv-mcp",
      "type": "http",
      "headers": {
        "Authorization": "Bearer ${input:develocity-access-key}"
      }
    }
  }
}

To add the MCP server in IntelliJ IDE using the GitHub Copilot plugin:

Open IntelliJ IDE.
Go to File > Settings (or IntelliJ IDEA > Settings on macOS).
In the left sidebar, navigate to Tools > GitHub Copilot.
Click on the MCP Servers tab or section.
Click the Configure button.
Paste the following JSON configuration into the configuration editor, replacing <base-url> with your Develocity or Develocity Reporting Kit URL.
Click OK or Apply to save your changes.

Consult the documentation for more details.

{
  "servers": {
    "drv-ai": {
      "type": "http",
      "url": "<base-url>/drv-mcp",
      "requestInit": {
        "headers": {
          "Authorization": "Bearer <Develocity Access Key>"
        }
      }
    }
  }
}

Restart IntelliJ after adding the configuration to ensure the configuration is updated.

Add the following JSON configuration to your Gemini settings file, typically located at ~/.gemini/settings.json. More details can be found in the documentation.


{
  "mcpServers": {
    "drv-ai": {
      "httpUrl":"<base-url>/drv-mcp",
      "headers": {
        "Authorization": "Bearer <Develocity Access Key>"
      }
    }
  }
}

Claude Desktop does not support connecting to MCP servers using an access key for authentication, which is required by Develocity Analytics MCP server.

To use the MCP server with Claude Desktop, you will need to configure a local MCP proxy that converts the streamable HTTP API into the STDIO protocol. This moves the authentication requirements to the proxy rather than the desktop app.

A popular proxy server that is frequently used for this purpose is mcp-proxy.

Gradle is not affiliated with or involved in the development of `mcp-proxy`.

Please ensure that any third-party software you install complies with your organization's security policies.

Once mcp-proxy is installed, you can then set up Claude Desktop by modifying its configuration file to launch the proxy. To find the location of this configuration file, open the settings panel in the Claude Desktop app then select Developer Edit Config.

This will open a file explorer window with the configuration file to edit. You can then paste the following settings, replacing <base-url> and <Develocity access key>.

{
    "mcpServers": {
        "drv": {
            "command": "mcp-proxy",
            "args": [
                "--transport",
                "streamablehttp",
                "--stateless",
                "<base-url>/drv-mcp"
            ],
            "env": {
                "API_ACCESS_TOKEN": "<Develocity access key>"
            }
        }
    }
}

Develocity Analytics MCP server follows the standard MCP protocol and will work with any clients that support:

Remote MCP servers
Streamable HTTP transport
MCP tools and prompts

Typically, a JSON configuration similar to the below can be used. Consult the documentation of your chosen application for more details on how to configure remote MCP servers.

{
  "servers": {
    "drv-ai": {
      "type": "http",
      "url": "<base-url>/drv-mcp",
      "headers": {
        "Authorization": "Bearer <Develocity Access Key>"
      }
    }
  }
}

Clients that only support stdio MCP servers may be configured to connect via the mcp-proxy tool as described in the #claude_desktop[Claude Desktop usage instructions].

Using the MCP server

Once the MCP server is configured in your application of choice, your selected LLM will be able to query your Develocity Reporting and Visualization data. For best results, we recommend that you prepare the model for this by using the prompt drv_insights_initial_prompt before asking your question.

Some examples of how to do this are outlined below in different client applications. As before, it is possible to use the MCP server with applications that are not listed, but the specific client has not been tested by Gradle. Please contact Develocity support if you encounter any issues.

Visual Studio Code
GitHub Copilot plugin for IntelliJ IDE
Gemini CLI
Claude Desktop

Make sure that Agent mode is selected in the chat. Prompts provided by the MCP server can be called via a slash command. The command is prefixed by the name of the MCP server used in the configuration.

Example of using the MCP prompt with Visual Studio Code

Once prepared this way, you may ask your question and the selected model will answer.

Example of using the MCP prompt with the IntelliJ Copilot Plugin

Once prepared this way, you may ask your question and the selected model will answer.

Prompts provided by the MCP server can be called via a slash command.

Example of using the MCP prompt with Gemini CLI

Once prepared this way, you may ask your question and the selected model will answer.

Claude Desktop does not have built-in support to use slash commands, so you need to select them in the UI.

To add the context prompt, you have to click the + sign below the text input and navigate to your configured MCP server.

Preparing to add the MCP prompt with Claude

Adding the MCP prompt with Claude

Select the prompt to add.

Selecting the MCP prompt with Claude Desktop

The prompt can be found using the search bar after pressing the + button

Searching for the MCP prompt with Claude Desktop

Recommended models

After testing the MCP server against a variety of different models, our recommendation is to use the latest version of your selected model where possible. Models such as GPT-5, Gemini 2.5 Pro, and Claude Sonnet 4.5 all provide good results.

Our testing has shown the "light" version of these models, such as GPT-5 Mini or Gemini 2.5 Flash are prone to making poor SQL query choices or failing to produce interesting responses to complex prompts. That said, they may still be able to offer useful basic analysis.

Avoiding common issues with LLMs

While provided MCP tools are deterministic in nature, the LLM’s effective use of them is not. For this reason, we recommend that you briefly inspect the SQL queries submitted by the LLM to the MCP server to catch obvious errors.

Long conversations with many different lines of questioning or analysis may lead to the model’s responses becoming gradually worse. Consider starting a new session or clearing the model’s context to fix this when it happens.

Authentication troubleshooting

Authentication or access key errors may be interpreted by client applications as a trigger to initiate the OAuth access flow. If you begin to see messages or errors related to authentication or OAuth, verify that your access key configuration is correct, the key is valid, and the associated Develocity user has the required permissions.

Some clients may need a restart for MCP configuration to be correctly updated.

Example prompts

Here are some example prompts to get started with once your AI application is connected to the MCP server:

"Are any projects using a version of log4j earlier than 2.17.1?"
"Which projects are using a non-LTS JDK version?"
"What tasks should I prioritize fixing to stabilize my CI builds?"
"Can you identify whether some users are impacted more by locally failing builds than others?"
"What’s the typical configuration time for local builds?"
"Which are the longest-running tasks in my CI builds?"
"What are the top 5 slowest builds in the last week?"

Difference with Develocity MCP server

Develocity Analytics MCP server is able to work in tandem with the Develocity MCP server. While Develocity Analytics MCP server described in this guide is aimed at analyzing larger batches of aggregate data, Develocity MCP server is aimed at analyzing individual builds and build failures. Some data, such as test data or failure grouping data, is only available from Develocity MCP server.

Both servers can work independently without the other being set up. However, for the best results, we recommended using both together.