Navigate:
Serena
~$SEREN1.6%

Serena: Coding agent toolkit with semantic capabilities

Coding agent toolkit with semantic code retrieval and editing for LLMs via MCP server.

LIVE RANKINGS • 02:49 PM • STEADY
TOP 10TOP 10TOP 10TOP 10TOP 10TOP 10TOP 10TOP 10TOP 10TOP 10TOP 10TOP 10
OVERALL
#9
DEVELOPER TOOLS
#1
1
30 DAY RANKING TREND
ovr#9
·Devel#1
STARS
20.3K
FORKS
1.4K
7D STARS
+320
7D FORKS
+15
See Repo:
Share:

Learn more about Serena

Serena is a coding agent toolkit that transforms LLMs into code-aware agents through semantic analysis and editing capabilities. The system operates through multiple integration methods including Model Context Protocol (MCP) servers, OpenAPI connections, and direct framework integration. It provides symbol-level code operations like find_symbol, find_referencing_symbols, and insert_after_symbol rather than basic text manipulation. The toolkit supports over 30 programming languages through LSP integration and JetBrains IDE plugin architecture.

Serena

1

Symbol-Level Operations

Performs code analysis and editing at the symbol level using semantic understanding rather than text-based search and replace operations.

2

Multiple Backend Support

Supports both Language Server Protocol (LSP) integration for 30+ languages and JetBrains IDE plugin architecture for enhanced code analysis.

3

Framework Agnostic

Provides decoupled tool implementations that can integrate with various LLMs, agent frameworks, and client interfaces through MCP, OpenAPI, or direct integration.


vv0.1.4

Likely the last release before stable 1.0.0, with extended CLI and simplified configuration.

  • Initial instructions no longer need to be loaded by the user
  • Significantly extended CLI
  • Removed replace_regex tool from ide-assistant and codex contexts
  • Simplify customization of modes and contexts, including CLI support
vv0.1.3

First PyPI release with improved stability by reducing asyncio usage and switching to synchronous components.

  • Reduce the use of asyncio to a minimum, improving stability and reducing the need for workarounds
  • Switch to newly developed fully synchronous LSP library solidlsp (derived from multilspy)
  • Switch from fastapi (which uses asyncio) to Flask in the Serena dashboard
  • The MCP server is the only asyncio-based component now, which resolves cross-component loop contamination
  • Improved editing tools: The editing logic was simplified and improved, making it more robust
v2025-05-19

Adds PHP support and fixes various small issues, though Java LS may not run on macOS in this version.

  • PHP support, fixed various small issues
  • Java LS doesn't seem to run on macos in this version

See how people are using Serena

Loading tweets...


[ EXPLORE MORE ]

Related Repositories

Discover similar tools and frameworks used by developers