A personal context store that gives AI coding agents like Codex, Claude Code and OpenCode persistent project knowledge via a CLI, MCP server and desktop app.
oraios/serena
An MCP server that gives coding agents IDE-level symbol search, semantic editing and refactoring over a codebase.
Open-source alternative to:JetBrains AI Assistant
- Stars
- 30,379
- Forks
- 2,232
- Open issues
- 101
- Last commit
- 5 days ago
What is it?
Serena is a toolkit that connects to AI coding agents through the Model Context Protocol (MCP) and gives them tools that work at the symbol level, such as finding a function, its references, or its implementations, instead of relying on line numbers and plain text search. It can be used with clients like Claude Code, Codex, Claude Desktop and IDE assistant plugins for VS Code, Cursor and JetBrains.
Serena offers two backends. By default it uses language servers implementing the Language Server Protocol, covering over 40 languages. Alternatively, a paid JetBrains plugin uses the analysis and refactoring engine of JetBrains IDEs, which adds move, inline and safe-delete operations and an interactive debugging tool.
The project also includes a memory system for sharing knowledge across agent sessions and a configuration layer written in YAML, with settings that can be scoped globally, per launch command, per project, or per client.
Key features
- 01Find symbols, references and implementations across a codebase.
- 02Replace, insert before or after symbol bodies without line-number edits.
- 03Rename symbols with cross-file reference updates.
- 04Supports over 40 languages through language servers.
- 05Optional JetBrains plugin adds move, inline and interactive debugging.
Use cases
- Letting a coding agent rename or move code safely in a large multi-file project.
- Giving an agent precise navigation in a monorepo with several languages.
- Keeping knowledge across agent sessions with the built-in memory system.
Quick start
- 1
Install uv (the only prerequisite)
- 2
Install Serena via uv
uv tool install -p 3.13 serena-agent - 3
Initialise Serena and verify the setup
serena init - 4
Configure a launch command in your MCP client (see the client guide)
Things to know
- The JetBrains backend requires a paid plugin, though a free trial is available.
- Language servers for some languages need extra dependencies installed separately.
- The license is GPL-3.0-or-later as shown in the badge, while the metadata reports NOASSERTION, so verify before commercial use.
Topics
FAQ
- What is serena?
- An MCP server that gives coding agents IDE-level symbol search, semantic editing and refactoring over a codebase. Serena is a toolkit that connects to AI coding agents through the Model Context Protocol (MCP) and gives them tools that work at the symbol level, such as finding a function, its references, or its implementations, instead of relying on line numbers and plain text search. It can be used with clients like Claude Code, Codex, Claude Desktop and IDE assistant plugins for VS Code, Cursor and JetBrains.
- Is serena free to use? What license does it have?
- serena is open source under the Other license. Check the license terms before using it in commercial products.
- Is serena still maintained?
- Actively maintained. The last commit to serena was on Oct 15, 2026.
- What can I use serena for?
- Letting a coding agent rename or move code safely in a large multi-file project. Giving an agent precise navigation in a monorepo with several languages. Keeping knowledge across agent sessions with the built-in memory system.
- How do I install serena?
- Install uv (the only prerequisite) Install Serena via uv: uv tool install -p 3.13 serena-agent Initialise Serena and verify the setup: serena init Configure a launch command in your MCP client (see the client guide)
- Where can I get serena?
- The source code and installation instructions are on GitHub: https://github.com/oraios/serena
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
A local code knowledge graph that gives AI coding tools precise review context over MCP and a CLI.
A hands-on Vibe Coding course that teaches non-programmers and developers to build real AI-assisted products from idea to launch.
Persistent memory for coding agents that captures session activity, compresses it with AI, and injects relevant context into future sessions.
MCP server for AI coding agents that keeps raw tool output out of the context window and persists session state across compaction.
A CLI that scans your codebase and generates compact AI context files (CLAUDE.md, .cursorrules, AGENTS.md) so coding assistants spend fewer tokens exploring the project.