JetBrains MCP Server: Connect an AI Assistant to Your IDE

JetBrains IDEs — including IntelliJ IDEA, PyCharm, WebStorm, and others — include a built-in MCP server that lets an AI assistant read and navigate the current project directly. Instead of pasting code into a chat, the assistant can call tools to read a file, search for a class, or inspect an error in the editor. Tools like fenbs extend this by keeping the project decisions and where-you-left-off notes outside the IDE session, so they are available when you switch IDEs, hand the work to a teammate, or start a fresh day.

The JetBrains MCP server ships as part of the IDE itself, enabled through settings. You do not need to install or configure a separate process. Connect it to an external AI client that supports MCP, and the assistant gains access to the tools the IDE exposes.

Enable the MCP server in your IDE

In recent JetBrains IDE versions, the MCP server is available in the AI settings. Enable it, and the IDE starts listening on a local port. The exact location varies by IDE version; check the JetBrains documentation for your specific version and IDE. Once enabled, you configure your external AI client — such as Claude Desktop or another MCP-compatible tool — to connect to the local endpoint.

The server runs only while the IDE is open. When you close the IDE, the server stops. This means a remote AI client that expects a persistent connection will lose access. Plan your workflow around this: use the IDE-connected session for code work, and keep the persistent project context in fenbs where it is available even when the IDE is closed.

What the JetBrains MCP server exposes

The tools available through the JetBrains MCP server typically include reading file content, searching for classes and methods by name, reading and navigating error messages in the editor, and retrieving project structure. The exact set of tools depends on the IDE version and which plugins are installed. Check the IDE's MCP settings for a list of enabled tools.

• Read a file by path relative to the project root.
• Search for a class, method, or symbol by name.
• List files in a directory.
• Read the current errors and warnings shown in the editor.
• Some IDEs expose run configuration and terminal access through additional tools.

What the IDE session cannot carry

An IDE session is temporary. When you close it, the session state disappears: what the assistant was working on, the decisions that were made during the session, and the next steps for the next developer. The JetBrains MCP server exposes the current file system state well, but it has no concept of project decisions, task progress, or handoff notes.

This is where fenbs adds value. Keep the tasks, decisions, and where-you-left-off notes in fenbs so they survive the IDE session. A teammate opening the project in a different IDE, or you returning to it after a few days, can read what was decided and what is next without relying on memory or old chat logs.

What the assistant can and cannot change

The JetBrains MCP server gives the assistant read access to files and the ability to navigate the project structure. Whether write operations are available depends on the specific tools the server exposes and which version of the IDE you are using. Check the IDE's MCP settings to see the full tool list and disable any write tools you do not need. An assistant that can read code is useful for explanation and review; an assistant that can write code changes files in your project, which requires the same judgment you would apply to any code change.

Combine IDE access with project context

Connect both the JetBrains MCP server and fenbs to your AI assistant. The IDE server gives the assistant access to the code; fenbs gives it access to the decisions and task state. A session that starts by reading "Where we left off" from fenbs and then reads the relevant file from the IDE is far more effective than one that starts from a blank context.

For team workflows, this also means a teammate using a different IDE — or even a different AI client — can continue from the same decisions and task state. The project context is not locked to one IDE or one session. See the Integrations page for the AI clients that connect to fenbs.

How to use JetBrains MCP with fenbs

Use fenbs for the durable project context and the IDE server for live code access. The two work together in the same AI session.

1. Create a free account and start a project for the codebase. Describe the goal, the technology stack, and the current work.
2. Write Decisions and rules for the project: which patterns to follow, which are deprecated, and which changes require review before merging.
3. At the end of each coding session, ask the assistant to save progress in fenbs: which tasks were completed, which are in progress, and what to start from next time.
4. Connect the JetBrains MCP server and fenbs to the same AI client. Ask the assistant to read the fenbs context before exploring the codebase.
5. When handing the work to a teammate, point them at the project in fenbs. They can read the current state without asking you to explain the history.