Serena MCP

Features

• Semantic code retrieval (symbol lookup, references, hierarchy)
• Symbolic code editing (rename, move, inline, delete)
• Code refactoring across projects and dependencies
• Language server protocol integration for broad language support
• JetBrains IDE integration for advanced debugging and analysis
• Memory system for long-lived agent workflows
• Configurable toolset and behavior

Use Cases

• Use when an AI agent needs to understand and modify complex codebases.
• Use when precise code refactoring is required, avoiding fragile text-based edits.
• Use to debug code interactively by setting breakpoints and inspecting variables.
• Use to discover and leverage code across project dependencies.

Non-Goals

• Executing arbitrary shell commands without explicit tool mapping.
• Replacing the core functionality of an IDE for human developers.
• Providing generic text-based search and replace when symbolic operations are applicable.

Workflow

1. Activate project and perform onboarding (if necessary).
2. Use `find_symbol` or `get_symbols_overview` to understand code structure.
3. Employ refactoring tools (`rename`, `move`, `inline`, `safe_delete`) for modifications.
4. Utilize `find_referencing_symbols` or `find_implementations` for code navigation.
5. Leverage `read_file` or `search_for_pattern` for broader text-based queries.
6. Use `execute_shell_command` for build or test execution when necessary.

Practices

• Symbolic code analysis
• Code refactoring
• Agent-driven debugging
• Configuration management