Skip to main content

Serena-Go - LSP-Powered Code Intelligence

Helix integrates with serena-go — a high-performance LSP (Language Server Protocol) based code analysis engine written in Go, providing deep code understanding capabilities for your programming assistant.

info

serena-go is a standalone project that AI Agent uses for LSP-powered code intelligence. It's a complete rewrite of the original Python Serena with better performance and simpler deployment.

✨ Core Features​

🔍 Symbol-Level Code Navigation​

  • Precise Search: Quickly locate classes, functions, and variable definitions
  • Reference Tracking: Find all usages of code across the project
  • Type Navigation: Jump to type definitions with one click
  • Hierarchy Analysis: Understand inheritance and call relationships

📝 Intelligent Code Editing​

  • Symbol-Level Modifications: Precisely replace function implementations without affecting surrounding code
  • Structure Awareness: Understand code structure to avoid syntax errors
  • Batch Refactoring: Smart symbol renaming with automatic reference updates
  • Context Completion: Intelligent suggestions based on project context

📊 Project-Level Understanding​

  • Architecture Insights: Quickly grasp overall project structure
  • Dependency Analysis: Understand inter-module dependencies
  • Monorepo Support: Automatic sub-project recognition
  • Cross-File Analysis: Understand code relationships across files

🚀 Technical Advantages​

30+ Programming Language Support​

serena-go supports mainstream programming languages through LSP:

PriorityLanguages
Tier 1Go, Python, TypeScript/JavaScript, Rust, Dart
Tier 2Java, C/C++, C#, Ruby, PHP
Tier 3Kotlin, Scala, Swift, Lua, Zig, Elixir, Haskell, Vue, Svelte, etc.

How it works: Reuses mature language server ecosystems (gopls, pyright, typescript-language-server, etc.) without reinventing the wheel.

Multi-Project Parallel Processing​

  • ✅ Analyze multiple projects simultaneously: Seamlessly switch between codebases
  • ✅ Monorepo auto-detection: Automatically discover and activate sub-projects
  • ✅ Independent LSP instances: Each project analyzed independently without interference

High-Performance Architecture​

  • ⚡ Native Go implementation: High performance, low memory footprint
  • 💾 Smart caching: Results cached after first analysis, sub-second subsequent queries
  • 🔄 Incremental updates: Only update affected parts when files change

💡 Use Cases​

1. Code Exploration​

"Where are the main entry points in this project?"
"List all API route definitions"
"Find all methods in the UserService class"

2. Code Modification​

"Rename getUserInfo function to fetchUserDetails"
"Replace the implementation logic of calculatePrice function"
"Add a new method to the User class"

3. Code Understanding​

"Explain how the processPayment function works"
"Where is this variable used?"
"Show the inheritance hierarchy of the Database class"

4. Refactoring Assistance​

"Find all calls to deprecated_function"
"Analyze dependencies of this module"
"Check for unused imports"

🎯 Using with AI Agent​

serena-go is built into the desktop application, ready to use out of the box:

  1. Open Workspace: Select project folder
  2. Auto Activation: serena-go automatically analyzes project structure
  3. Start Chatting: Directly ask code-related questions

Status Indicators:

  • 🟢 Green: serena-go ready
  • 🟡 Yellow: Analyzing project
  • 🔴 Red: Language server installation needed

Server Mode​

When running AI Agent backend service, serena-go starts automatically with workspace:

# Start AI Agent
./agentbackend --data-dir ~/.aiagent

# serena-go starts automatically with workspace
# No additional configuration needed

🔧 Language Servers​

serena-go requires corresponding LSP servers to analyze specific languages. Below is a comprehensive list of supported language servers with installation instructions.

Mainstream Languages​

LanguageLSP ServerInstallationHomepage
Gogoplsgo install golang.org/x/tools/gopls@latestgopls.dev
Pythonpyrightnpm install -g pyrightpyright
Pythonpylsp (alternative)pip install python-lsp-serverpylsp
TypeScript/JavaScripttypescript-language-servernpm install -g typescript-language-server typescriptts-ls
Rustrust-analyzerrustup component add rust-analyzerrust-analyzer
JavajdtlsDownload Eclipse JDT LSjdtls
C/C++clangdbrew install llvm (macOS) or Downloadclangd
C#csharp-lsdotnet tool install --global csharp-lscsharp-ls
C#OmniSharp (alternative)DownloadOmniSharp

Web & Frontend​

LanguageLSP ServerInstallationHomepage
HTMLvscode-html-languageservernpm install -g vscode-langservers-extractedvscode-html
CSS/SCSSvscode-css-languageservernpm install -g vscode-langservers-extractedvscode-css
JSONvscode-json-languageservernpm install -g vscode-langservers-extractedvscode-json
Vuevolarnpm install -g @vue/language-serverVolar
Sveltesvelte-language-servernpm install -g svelte-language-serversvelte-ls

Mobile & Cross-Platform​

LanguageLSP ServerInstallationHomepage
Dart/FlutterdartIncluded with Dart SDKDart
Kotlinkotlin-language-serverDownloadkotlin-ls
Swiftsourcekit-lspIncluded with Xcode or Swift Toolchainsourcekit-lsp

Systems & Low-Level​

LanguageLSP ServerInstallationHomepage
Zigzlsbrew install zls (macOS) or Build from sourcezls
Haskellhaskell-language-serverghcup install hlshls
OCamlocaml-lspopam install ocaml-lsp-serverocaml-lsp
Elixirelixir-lsDownloadelixir-ls

Scripting & Dynamic Languages​

LanguageLSP ServerInstallationHomepage
Rubysolargraphgem install solargraphsolargraph
PHPintelephensenpm install -g intelephenseintelephense
Lualua- install lua-language-server` (macOS)lua-ls
Perlperlnavigatornpm install -g perlnavigator-serverperlnavigator
Bashbash-language-servernpm install -g bash-language-serverbash-ls

Configuration & Markup​

LanguageLSP ServerInstallationHomepage
YAMLyaml-language-servernpm install -g yaml-language-serveryaml-ls
TOMLtaplocargo install taplo-cli --features lsptaplo
MarkdownmarksmanDownloadmarksman
LaTeXtexlabcargo install texlabtexlab

Data & Query Languages​

LanguageLSP ServerInstallationHomepage
SQLsqlsgo install github.com/lighttiger2505/sqls@latestsqls
GraphQLgraphql-language-servicenpm install -g graphql-language-service-cligraphql-ls

Installation Tips​

macOS​

# Install via Homebrew (recommended)
brew install llvm          # For clangd (C/C++)
brew install lua-language-server
brew install zls           # For Zig

# Install via npm (Node.js required)
npm install -g typescript-language-server typescript
npm install -g pyright
npm install -g vscode-langservers-extracted  # HTML, CSS, JSON
npm install -g yaml-language-server
npm install -g bash-language-server
npm install -g intelephense
npm install -g svelte-language-server
npm install -g @vue/language-server

# Install via language-specific package managers
go install golang.org/x/tools/gopls@latest
cargo install taplo-cli --features lsp
pip install python-lsp-server
gem install solargraph

Linux (Ubuntu/Debian)​

# Install system packages
sudo apt install clangd  # C/C++

# Install via npm
npm install -g typescript-language-server typescript
npm install -g pyright
npm install -g vscode-langservers-extracted
npm install -g yaml-language-server
npm install -g bash-language-server

# Install via snap
sudo snap install rust-analyzer --beta

# Language-specific
go install golang.org/x/tools/gopls@latest
pip install python-lsp-server
gem install solargraph

Windows​

# Install via Scoop (recommended)
scoop install llvm        # For clangd
scoop install lua-language-server

# Install via npm
npm install -g typescript-language-server typescript
npm install -g pyright
npm install -g vscode-langservers-extracted
npm install -g yaml-language-server
npm install -g bash-language-server

# Install via Chocolatey
choco install rust-analyzer

# Language-specific
go install golang.org/x/tools/gopls@latest
pip install python-lsp-server

Verification​

After installation, verify language servers are available:

# Check if language server is in PATH
which gopls                    # Go
which pyright                  # Python
which typescript-language-server  # TypeScript
which rust-analyzer            # Rust
which clangd                   # C/C++

# Test language server
gopls version
pyright --version
typescript-language-server --version

Configuration​

Most language servers work out-of-the-box with serena-go. For advanced configuration:

Go (gopls)

// .gopls.json
{
  "ui.diagnostic.analyses": {
    "unusedparams": true,
    "shadow": true
  },
  "ui.completion.usePlaceholders": true
}

Python (pyright)

// pyrightconfig.json
{
  "typeCheckingMode": "basic",
  "reportMissingImports": true,
  "reportMissingTypeStubs": false
}

TypeScript

// tsconfig.json (in project root)
{
  "compilerOptions": {
    "strict": true,
    "esModuleInterop": true
  }
}

C/C++ (clangd)

# .clangd
CompileFlags:
  Add: [-std=c++17, -Wall]
  Remove: [-W*]
Auto-Installation

Helix will detect when a language server is missing and prompt you with installation instructions. You can also configure automatic installation in Settings → Language Servers → Auto-install.

Project-Specific Requirements

Some language servers require project-specific files to work properly:

  • Go: go.mod file
  • TypeScript: tsconfig.json or package.json
  • Python: pyproject.toml, setup.py, or requirements.txt
  • Rust: Cargo.toml
  • Java: pom.xml or build.gradle
  • C/C++: compile_commands.json or .clangd

serena-go will automatically detect these files and configure language servers accordingly.

📈 Performance Characteristics​

Initial Analysis​

  • Small projects (< 100 files): < 5 seconds
  • Medium projects (100-1000 files): 5-30 seconds
  • Large projects (> 1000 files): 30-60 seconds

Subsequent Operations​

  • Symbol lookup: < 100ms
  • Reference search: < 500ms
  • Code navigation: instant

Resource Usage​

  • Memory: 50-200MB per project
  • CPU: Peak < 50% during analysis, < 5% when idle

🆚 Comparison with Traditional Methods​

Featureserena-go (LSP)Traditional Text Analysis
Precision🟢 Semantic understanding🔴 Regex matching
Refactoring Safety🟢 Auto-update references🔴 Manual find/replace
Cross-File Analysis🟢 Native support🟡 Partial support
Language Support🟢 30+ languages🟡 Limited
Performance🟢 Cache optimized🟡 Re-scan each time

💬 Real-World Examples​

Example 1: Quick Familiarization with New Project​

👤 "What does this project do? What are the main modules?"

🤖 [Using serena-go to analyze]
   - Entry file found: cmd/main.go
   - Core packages identified: pkg/api, pkg/database, pkg/service
   - API endpoints listed: 12 REST interfaces
   - Data models: 5 main entities

Example 2: Safe Refactoring​

👤 "Rename GetUser to FetchUser, update all calls"

🤖 [Using serena-go to execute]
   ✅ Function definition renamed
   ✅ 23 call sites updated
   ✅ 5 test files updated
   ✅ Type safety maintained

Example 3: Understanding Complex Logic​

👤 "Explain the workflow of processOrder function"

🤖 [Using serena-go to analyze]
   1. Read function implementation → understand control flow
   2. Trace called sub-functions → build call graph
   3. Analyze data flow → understand state changes
   4. Generate natural language explanation

🔮 Future Features​

  • 🚧 Real-time Error Detection: Instant feedback while editing
  • 🚧 Intelligent Completion: Context-aware code suggestions
  • 🚧 Test Generation: Auto-generate tests from function signatures
  • 🚧 Documentation Generation: Extract comments to generate docs

🔗 Related Resources​

Note: serena-go is AI Agent's built-in MCP server, no additional installation required. Through LSP protocol, it enables AI to truly "understand" your code, not just process text.