Gene Library Courses Download Pricing Contact Sign in
jetbrains logo
developer-tools Official Website

JetBrains MCP Server – IDE Code Intelligence

JetBrains' built-in MCP Server lets external AI clients interact with IntelliJ-based IDEs such as IntelliJ IDEA, PyCharm, WebStorm, PhpStorm, Rider, and Android Studio. Use it when an agent needs IDE-aware code analysis, project context, build feedback, file search, debugging, database inspection, or controlled execution inside an open JetBrains project.

#ide#code-analysis#debugging

Overview

JetBrains provides an integrated MCP Server for IntelliJ-based IDEs. It lets
external AI clients such as Claude Desktop, Claude Code, Cursor, Codex, VS Code,
Windsurf, and other MCP-compatible tools access IDE-provided project context
without leaving the user's development environment.

What the MCP server enables

The JetBrains MCP Server exposes IDE tools that can be enabled or disabled in
Settings | Tools | MCP Server | Exposed Tools. Depending on the IDE, installed
plugins, and selected tools, an AI agent can:

  • Build the project or selected files and retrieve compilation errors.
  • Analyze files with JetBrains inspections.
  • List modules and project dependencies.
  • Retrieve symbol information using IDE code insight.
  • Search project files by text or regular expression.
  • Modify files in the project when write tools are exposed.
  • Run configurations and terminal commands.
  • Inspect VCS roots.
  • Access database connection metadata and schema details when Database Tools
    and SQL support is installed and enabled.
  • Use debugger tools when the Debugger MCP toolset is available.

When to use it

Use JetBrains MCP when an AI agent needs real IDE context rather than generic
filesystem access. Practical examples include asking an assistant to locate a
symbol, inspect compilation errors after an edit, find usages by search,
validate a file with inspections, summarize project modules, inspect a database
schema from the IDE data source, or run a test configuration with user approval.

Connection and authentication

In JetBrains IDEs version 2025.2 and later, the MCP Server plugin is bundled and
enabled by default. Enable the server in Settings | Tools | MCP Server, then use
Auto-Configure for supported clients or Copy SSE Config, Copy Stdio Config, or
Copy HTTP Stream Config for manual setup. The exact local endpoint or command is
generated by the IDE because it depends on the running IDE instance and local
web server settings.

The deprecated @jetbrains/mcp-proxy package previously provided stdio
bridging, but JetBrains states that this repository is no longer maintained
because the core functionality is now integrated into IntelliJ-based IDEs.

Key considerations

JetBrains MCP can access project files and can execute terminal commands or run
configurations when those tools are enabled. By default, command execution is
confirmation-oriented; JetBrains provides a separate “brave mode” setting that
allows shell commands or run configurations without confirmation. Enable that
only for trusted projects and trusted AI clients. Database tools require the
relevant JetBrains plugins, and read-only database safety should be enforced by
using a database account with read-only privileges. Review tool exposure in the
IDE settings and disable tools that are unnecessary for the workflow.

Supported Transports

streamable_http

URL: http://127.0.0.1:${JETBRAINS_MCP_PORT}/mcp

sse

URL: http://127.0.0.1:${JETBRAINS_MCP_PORT}/sse

stdio

Command: ${JETBRAINS_MCP_STDIO_COMMAND}

Args:

  • ${JETBRAINS_MCP_STDIO_ARG_1}
  • ${JETBRAINS_MCP_STDIO_ARG_2}

Frequently Asked Questions

When should an AI agent use the JetBrains MCP Server?
Use it when an agent needs IDE-aware project context, such as inspecting symbols, finding code, running inspections, building files, checking project modules, reviewing dependencies, running test configurations, or using debugger and database tools exposed by a JetBrains IDE.
What does the JetBrains MCP Server add to an AI agent's capabilities?
It gives the agent access to JetBrains IDE intelligence, including code insight, inspections, project structure, build feedback, file search, VCS roots, database metadata, debugger tooling, and controlled execution workflows that generic filesystem access cannot provide.
What can an AI agent access or manage through JetBrains MCP?
Depending on enabled exposed tools, the agent can analyze files, retrieve symbol information, build the project, search files, modify project files, run configurations, execute terminal commands, inspect VCS roots, and access database or debugger features when the relevant IDE plugins are available.
How is authentication configured for the JetBrains MCP Server?
JetBrains MCP is a local IDE integration rather than a hosted account API. Enable the MCP Server in the IDE and configure trusted local clients from Settings | Tools | MCP Server. Security is controlled by local client access, tool exposure settings, confirmation prompts, and any database or project permissions configured in the IDE.
Which transport should be used for the JetBrains MCP Server?
Use the transport copied from Settings | Tools | MCP Server for the running IDE. JetBrains documents SSE, stdio, and HTTP Stream configuration options. Prefer the IDE-generated configuration, because endpoint ports and stdio commands are local to the running JetBrains IDE instance.