Features

Ctxman is a powerful AI Development Platform with a comprehensive set of features designed for modern development workflows.

🚀 Platform Features

Plugin Architecture

Ctxman features a modular plugin system that makes it easy to extend and customize:

• Language Plugins: Add support for new programming languages
• Exporter Plugins: Create custom output formats
• Analyzer Plugins: Extend analysis capabilities
• Easy Integration: Simple API for creating plugins

// Example plugin structure
export default {
  name: 'my-plugin',
  type: 'language',
  extensions: ['.xyz'],
  analyze: async (content, filePath) => {
    // Custom analysis logic
  },
};

Git Integration

Seamlessly integrate with your Git workflow:

• Changed Files Analysis: Analyze only modified files
• Diff Analysis: Compare changes between commits
• Author Tracking: Track code contributions
• Branch Comparison: Compare branches

# Analyze changed files in the last commit
ctxman --git-diff HEAD~1

# Analyze files changed in a specific branch
ctxman --git-branch feature/new-feature

Watch Mode

Real-time file monitoring for instant feedback:

• Auto-Analysis: Automatically re-analyze on file changes
• Live Updates: See results instantly
• Configurable Intervals: Set custom debounce timing
• Selective Watching: Watch specific directories

# Start watch mode
ctxman watch

# Watch with specific options
ctxman watch --interval 1000 --output json

REST API

HTTP server for programmatic access:

• 6 Endpoints: Full API coverage
• JSON Responses: Standard REST format
• CORS Support: Cross-origin requests
• Health Check: Monitor server status

# Start API server
ctxman serve

# API is available at http://localhost:3000
curl http://localhost:3000/api/analyze

🎨 User Interface

Interactive Wizard Mode

User-friendly guided setup for beginners:

• Step-by-Step Process: Clear guidance through options
• Smart Defaults: Sensible default values
• Input Validation: Catch errors early
• Help Text: Contextual help at every step

CLI Mode

Traditional command-line interface for power users:

• Direct Commands: Skip the wizard
• Shell Integration: Perfect for scripts
• Exit Codes: Proper error handling
• Piping Support: Chain with other tools

🔢 Token Analysis

Exact Token Counting

Accurate token counting for LLM planning:

• Per model: --target-model picks the encoding (o200k_base for current OpenAI models, cl100k_base otherwise); counts for models with their own tokenizer are marked approximate
• Exact for Claude: ctxman count --model claude-opus-5-5 --api asks Anthropic's count_tokens API (free; secrets are redacted first)
• Current models: ctxman --list-llms lists windows, output limits and prices of 40+ models
• Per-File Analysis: Detailed breakdown

Multi-Language Support

Analyze code in 14+ programming languages:

Language	Extensions	Features
JavaScript	.js, .jsx, .mjs	Full AST analysis
TypeScript	.ts, .tsx	Type-aware analysis
Python	.py	Method-level analysis
PHP	.php	Class and function analysis
Ruby	.rb	Module and method analysis
Java	.java	Class and method analysis
Kotlin	.kt	Function analysis
C#	.cs	Class and method analysis
Go	.go	Function analysis
Rust	.rs	Function and struct analysis
Swift	.swift	Function analysis
C/C++	.c, .cpp, .h	Function analysis
Scala	.scala	Class and function analysis

Method-Level Analysis

Get granular insights into your code:

• Per-Function Tokens: Token count for each function
• Complexity Metrics: Understand code complexity
• Largest Functions: Identify token-heavy code
• Optimization Tips: Suggestions for reduction

🎯 Filtering & Configuration

Dual Ignore System

Flexible file filtering:

• .gitignore Support: Respect Git ignore rules
• .contextignore: Custom ignore patterns
• .contextinclude: Include-only mode
• Priority System: Clear precedence rules

Method Filtering

Granular control over analysis:

• .methodinclude: Include specific methods
• .methodignore: Exclude specific methods
• Pattern Matching: Wildcards and regex support
• Performance: Skip unnecessary analysis

📊 Output Formats

TOON Format

40-50% token reduction with optimized output:

• Compact Structure: Minimized redundancy
• Semantic Preservation: Keep meaning intact
• LLM Optimized: Perfect for AI consumption
• Backward Compatible: Easy to convert back

GitIngest Format

GitHub-ready output format:

• Markdown Compatible: Works with GitHub
• Syntax Highlighting: Proper code blocks
• File Trees: Visual directory structure
• Line Numbers: Reference specific lines

Standard Formats

Multiple output options:

• JSON: Machine-readable
• Markdown: Documentation-ready
• Plain Text: Simple and clean
• Custom: Build your own with plugins

⚡ Performance

Caching System

Speed up repeated analyses:

• Intelligent Cache: Only re-analyze changed files
• Configurable TTL: Set cache expiration
• Cache Invalidation: Manual or automatic
• Memory Efficient: Low overhead

Measured Speed

Handle large codebases efficiently (4-vCPU Linux VM, Node.js 22):

• Fast Scanning: ~1,900 files discovered with ignore rules in ~75 ms
• Token Counting: 1,135 files / 2.6M tokens counted with tiktoken in ~6 s
• Progress Reporting: Track analysis progress

🔧 Configuration

Flexible Configuration

Multiple configuration options:

• JSON Config: Structured configuration
• CLI Flags: Quick overrides
• Environment Variables: CI/CD integration
• Defaults: Sensible out-of-the-box behavior

Example Configuration

{
  "output": {
    "format": "toon",
    "file": "output.md"
  },
  "analysis": {
    "methodLevel": true,
    "tokenCount": true
  },
  "exclude": ["node_modules", "dist", "*.test.js"]
}

🔐 Security

Safe by Default

Security-first approach:

• No External Calls: All processing is local
• File Validation: Safe file handling
• Memory Limits: Prevent resource exhaustion
• Input Sanitization: Clean all inputs

---

Ready to get started? Check out our Getting Started Guide or try the Interactive Demo.