Serena

v1.0.0Coding Agentsstable

A coding agent toolkit that provides IDE-like semantic code retrieval and editing tools, enabling LLMs to efficiently navigate and modify codebases using symbol-level operations instead of basic file reading and string replacements.

agentaiai-codingclaudeclaude-code
Share:
24,484
Stars
0
Downloads
0
Weekly
0/5

What is Serena?

Serena is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to coding agent toolkit that provides ide-like semantic code retrieval and editing tools, enabling llms to efficiently navigate and modify codebases using symbol-level operations instead of basic file re...

A coding agent toolkit that provides IDE-like semantic code retrieval and editing tools, enabling LLMs to efficiently navigate and modify codebases using symbol-level operations instead of basic file reading and string replacements.

This server falls under the Coding Agents category on MCPgee, the world's largest MCP server directory with 33,000+ servers.

Features

  • A coding agent toolkit that provides IDE-like semantic code

Use Cases

Semantic code retrieval
Symbol-level editing
Codebase navigation
oraios

Maintainer

LicenseMIT
Languagepython
Versionv1.0.0
UpdatedMay 22, 2026
Statushealthy
Maintenanceactive

Works with

ClaudeOpenAIwindowsmacoslinux

Installation

NPM

npx -y serena

Manual Installation

npx -y serena

Configuration

Configuration Details

Config File

claude_desktop_config.json

Performance

Response Metrics

Response Time< 200ms
ThroughputMedium

Resource Usage

Memory UsageLow
CPU UsageLow

How to Set Up and Use Serena

Serena is an IDE-grade coding agent toolkit that provides semantic code retrieval, editing, refactoring, and debugging tools for LLMs, operating at the symbol level rather than relying on raw line numbers or string replacements. It integrates with any MCP-compatible AI client — including Claude Code, Cursor, Copilot, and JetBrains AI Assistant — and supports over 40 programming languages through Language Server Protocol (LSP) backends or the paid Serena JetBrains Plugin. By exposing tools like find-symbol, find-references, rename, replace-symbol-body, and cross-file move, Serena lets agents navigate and modify large, complex codebases with significantly fewer tokens and far fewer errors than approaches based on file reading and text search. It is the recommended MCP toolkit when AI agents need to work reliably on real production codebases.

Prerequisites

  • Python 3.11 or later installed (Serena's core is Python-based)
  • A language server for your target language (e.g., pylsp for Python, typescript-language-server for TypeScript) installed and accessible in PATH
  • An MCP-compatible client such as Claude Code, Claude Desktop, Cursor, or a JetBrains IDE with AI Assistant
  • Node.js 18+ if installing via the npx method
  • Optional: a JetBrains IDE with the Serena Plugin for enhanced refactoring and debugging capabilities
1

Install Serena following the official Quick Start (not via marketplaces)

The Serena team explicitly warns against installing from MCP or plugin marketplaces because those listings contain outdated commands. Install directly from the GitHub repository using uvx or pip to get the correct, up-to-date version.

# Using uvx (recommended)
uvx serena@latest

# Or install with pip
pip install serena
2

Install a language server for your primary language

Serena uses LSP-compatible language servers to provide semantic understanding. Install the server for your project's language and ensure it is on your PATH. Examples below cover Python and TypeScript.

# Python
pip install python-lsp-server

# TypeScript / JavaScript
npm install -g typescript-language-server typescript

# Go
go install golang.org/x/tools/gopls@latest
3

Add the Serena MCP server to your client configuration

Configure the Serena MCP server in your MCP client's config file. You must pass the project directory as an argument so Serena knows which codebase to index and provide semantic tools for.

4

Restart your MCP client and open your project

After saving the config and restarting the client, Serena indexes the specified project directory via the language server. The first indexing may take a few seconds for large codebases. Once complete, your AI agent gains access to all of Serena's semantic tools.

5

Use Serena tools through your AI agent

Ask your AI assistant to navigate, refactor, or edit code using Serena's symbol-level capabilities. The agent will automatically call tools like find_symbol, find_referencing_symbols, replace_symbol_body, and safe_delete instead of reading entire files.

Serena Examples

Client configuration

Add this block to your claude_desktop_config.json. Replace /path/to/your/project with the absolute path to the project Serena should analyze. The project_path argument is required.

{
  "mcpServers": {
    "serena": {
      "command": "npx",
      "args": ["-y", "serena", "--project-path", "/path/to/your/project"]
    }
  }
}

Prompts to try

Once Serena is connected, your AI agent can use these to perform symbol-level operations on your codebase.

- "Find all callers of the process_payment function and list them with their file locations"
- "Rename the class UserService to AccountService across the entire codebase"
- "Replace the body of the calculate_discount method with this new implementation: ..."
- "Find all implementations of the IRepository interface and check if they implement the new save() method"
- "Safe-delete the unused helper function format_legacy_date"

Troubleshooting Serena

Serena reports 'no language server found' or tools return empty results

Ensure the language server for your project's language is installed and available in PATH. Test by running 'pylsp --version' (Python) or 'typescript-language-server --version' (TypeScript) in a terminal. Serena auto-detects language servers based on file extensions in your project.

Symbol lookups return stale or incorrect results after code changes

Serena's language server connection maintains a live index but may need a moment to re-index after large changes. If results are consistently wrong, restart the MCP server (restart your client) to force re-initialization of the language server connection.

Serena's tools are disabled when used inside Claude Code or Codex

When Serena is used inside an agentic harness like Claude Code, basic file tools (read_file, list_dir, execute_shell_command) are disabled by default to avoid overlap with the harness's built-in tools. This is expected behavior — the semantic tools (find_symbol, rename, replace_symbol_body) remain fully active.

Frequently Asked Questions about Serena

What is Serena?

Serena is a Model Context Protocol (MCP) server that coding agent toolkit that provides ide-like semantic code retrieval and editing tools, enabling llms to efficiently navigate and modify codebases using symbol-level operations instead of basic file reading and string replacements. It connects AI assistants to external tools and data sources through a standardized interface.

How do I install Serena?

Install via npm with the command: npx -y serena. Then add the server configuration to your AI client's JSON config file (e.g., claude_desktop_config.json or .cursor/mcp.json).

Which AI clients work with Serena?

Serena works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.

Is Serena free to use?

Yes, Serena is open source and available under the MIT license. You can use it freely in both personal and commercial projects.

Browse More Coding Agents MCP Servers

Explore all coding agents servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.

Quick Config Preview

{ "mcpServers": { "serena": { "command": "npx", "args": ["-y", "serena"] } } }

Add this to your claude_desktop_config.json or .cursor/mcp.json

Read the full setup guide →

Ready to use Serena?

Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.

33,000+ ServersFree & Open SourceStep-by-Step Guides