Coda's Tree-Sitter Intelligence provides comprehensive code intelligence capabilities using tree-sitter for parsing and analysis. It enables AI-powered code understanding, navigation, and analysis across 30+ programming languages.

1. Code Analysis: Extract definitions, imports, and references from source files
2. Multi-Language Support: 30+ languages including Python, JavaScript, Java, Kotlin, PHP, and more
3. Dependency Graphs: Visualize and analyze code dependencies with cycle detection
4. Hybrid Tools: Both native (fast) and MCP (comprehensive) tool implementations
5. Smart Navigation: Find definitions, analyze files, and understand code structure

# Intelligence commands
coda /intel analyze <file>     # Analyze a file's structure
coda /intel find <name>        # Find definitions by name
coda /intel deps <file>        # Show file dependencies
coda /intel scan <dir>         # Scan directory for code
coda /intel graph <dir>        # Build dependency graph
coda /intel map [path]         # Map repository structure
coda /intel stats              # Show language statistics

The AI assistant understands code-related queries:

• "Find the definition of MyClass"
• "What does this file import?"
• "Show me all functions in this directory"
• "Analyze the structure of main.py"

Verify the intelligence tools are available:

coda /intel help

coda chat

# Then ask:
> "Find the UserService class"
> "What's in the utils.py file?"
> "Show me code statistics for src/"

# Find a definition
coda /intel find TreeSitterAnalyzer

# Analyze a file
coda /intel analyze coda/intelligence/tree_sitter_analyzer.py

# Scan a directory
coda /intel scan coda/intelligence/

# Build dependency graph
coda /intel graph src/ --output=dependencies.json

# Native tools (simple & fast)
from coda.agents.intelligence_tools import find_definition
print(find_definition("MyClass"))

# MCP tools (comprehensive)
from coda.tools.intelligence_tools import AnalyzeFileTool
tool = AnalyzeFileTool()
result = await tool.execute({"file_path": "main.py"})

coda/
├── intelligence/              # Core intelligence module
│   ├── tree_sitter_analyzer.py
│   ├── repo_map.py
│   ├── dependency_graph.py
│   └── queries/              # Language query files
│       ├── python-tags.scm
│       ├── javascript-tags.scm
│       └── ... (30+ languages)
├── tools/
│   └── intelligence_tools.py  # MCP tool implementations
└── agents/
    └── intelligence_tools.py  # Native tool implementations

Run the test suite:

# All intelligence tests
pytest tests/intelligence/

# Specific test files
pytest tests/intelligence/test_tree_sitter_analyzer.py
pytest tests/tools/test_intelligence_tools.py
pytest tests/agents/test_intelligence_tools.py

• Design Document - Architecture and implementation details
• Usage Guide - Comprehensive usage examples
• Examples - Real-world usage examples

1. Create a query file: coda/intelligence/queries/[language]-tags.scm
2. Add extension mapping in tree_sitter_query_analyzer.py
3. Write tests in tests/intelligence/test_[language]_support.py
4. Update documentation

;; Class definitions
(class_declaration
  name: (identifier) @name.definition.class) @definition.class

;; Function definitions
(function_declaration
  name: (identifier) @name.definition.function) @definition.function

1. PL/SQL: Limited support due to tree-sitter grammar constraints
2. Large Files: Performance may degrade on very large files (>10k lines)
3. Cross-File Resolution: Currently analyzes files independently

• Check if file extension is mapped in _get_extension_map()
• Ensure corresponding query file exists
• May fall back to regex-based parsing

• Use native tools for quick queries
• Limit directory scanning depth with recursive=False
• Filter by file patterns when scanning

• Verify query file captures import patterns correctly
• Check for language-specific import syntax
• Some languages may have limited import detection

• Incremental parsing for better performance
• Cross-file symbol resolution
• Type information extraction
• Semantic code search
• IDE integration via Language Server Protocol

Please report issues or suggestions:

• Create an issue with [tree-sitter] prefix
• Include language and code sample for parsing issues
• Suggest new languages or features

Same as Coda project - see main LICENSE file.