Skip to content
Serena

SerenaMCP Server

Added to Onei
4 views
MIT

Summary

Semantic code toolkit for coding agents — symbol-level search, editing and refactoring backed by language servers or the JetBrains platform.

Features

  • Symbol-level code search across a whole project
  • Find all references to a symbol before changing it
  • Insert, replace and rename symbols as atomic edits
  • Language Server Protocol backend covering many languages
  • Optional JetBrains IDE backend for large codebases
  • Reads only the symbols an agent needs, saving context
  • Project-based workflow with per-project configuration

Installation

Set up this MCP server in your favorite AI agent — copy a ready-made configuration below.

Any MCP-compatible agent

Most agents (Claude, Cursor, Windsurf, VS Code, and more) read a standard mcpServers configuration.

  1. Open your agent's MCP configuration file.
  2. Merge the snippet below into it.
  3. Restart the agent — the "Serena" tools will be available.
{
  "mcpServers": {
    "serena-mcp-server": {
      "command": "serena",
      "args": [
        "start-mcp-server",
        "--context=desktop-app"
      ]
    }
  }
}

Description

Serena gives a coding agent the tools an IDE gives a developer. Rather than grepping for strings and rewriting files by line number, an agent connected to Serena works at the level of symbols: find a class or function by name, list every reference to it, read only the body it needs, insert a method after another one, or rename a symbol across the whole project as a single atomic operation.

The semantic layer comes from real language tooling. By default Serena runs Language Server Protocol servers — the same ones that power editor autocomplete — which keeps it free and open source across a wide range of languages. Alternatively it can be driven by a JetBrains IDE backend, which trades the extra setup for the deeper indexing JetBrains does on large codebases.

The practical payoff is token efficiency on big repositories. Loading a 4,000-line file to change one method wastes most of the context window; asking for that one symbol does not. This matters most where naive agents struggle: multi-file renames, moves, and reference lookups that would otherwise take a dozen careful, error-prone steps.

Serena is MIT-licensed, written in Python and distributed as the serena-agent tool via uv. It is a stdio MCP server, so it works with Claude Code, Claude Desktop, Codex and any MCP-capable IDE. The maintainers explicitly ask you to install from source rather than from plugin marketplaces, which carry outdated launch commands. Note that setup is a two-step process — install the tool, then run serena init — and that some languages need their language server installed separately.

Related MCP Servers

MCP: Hugging Face

by Hugging Face

New

Hugging Face's official MCP server — search the Hub's models, datasets and Spaces, read repository files, and call thousands of Gradio applications as tools from one remote endpoint.

New

Qase test management over MCP: 36 task-oriented tools for cases, runs, results, defects and QQL search, hosted at mcp.qase.io or self-run with your own API token.

New

Arm's official MCP server for porting and tuning software on Arm — semantic docs search, codebase migration scanning, container architecture checks and LLVM-MCA assembly analysis.

New

Official AntV MCP server that turns a description of your data into a rendered chart — 26 visualization types from bar and line to sankey, mind-map, fishbone and geographic maps.

Browse all MCP servers →