Arc Flow

AI-powered code intelligence and visualization for VS Code

Transform your codebase into an interactive knowledge graph with advanced dependency analysis and architectural insights.

Overview

Arc Flow helps you understand complex codebases through interactive visualizations and AI-powered insights. See dependencies, track technical debt, and navigate your code with confidence.

Features

🗺️ Interactive Code Graph

Visualize your entire codebase structure with three view modes:

• Architecture View - High-level domain structure
• Codebase View - Detailed file and symbol relationships
• Trace View - Function call chains and critical paths

🤖 AI-Powered Analysis

Get instant insights about your code:

• Code explanations and documentation
• Refactoring suggestions
• Security vulnerability analysis
• Risk assessment and impact analysis

📊 Technical Debt Detection

Automatically identify code quality issues:

• Complexity hotspots (cyclomatic complexity)
• Coupling issues between modules
• Change impact analysis (blast radius)
• Health scoring per domain (Healthy/Warning/Critical)

⚡ Real-Time Updates

• Incremental indexing - Only re-parses changed files
• File watcher integration - Auto-updates on save
• Fast symbol search - <10ms query time
• Scalable - Handles codebases with 50,000+ files

🔍 Advanced Features

• Multi-language support (TypeScript, Python, C/C++)
• Export graph data as JSON
• Symbol navigation with CodeLens
• Dependency visualization
• AI-powered architecture refinement

Installation

1. Open VS Code Extensions (Ctrl+Shift+X / Cmd+Shift+X)
2. Search for "Arc Flow"
3. Click Install

Quick Start

Step 1: Index Your Workspace

Open Command Palette (Ctrl+Shift+P / Cmd+Shift+P) and run:

Arc Flow: Index Workspace

Step 2: Visualize the Graph

Arc Flow: Visualize Code Graph

Step 3: Configure AI (Optional)

Arc Flow: Configure AI Keys

Supported AI Providers:

• Groq (Llama 3.1) - Ultra-fast analysis (<300ms)
• Google Gemini 1.5 Pro - Deep architectural analysis
• AWS Bedrock (Amazon Nova 2) - Enterprise-grade AI

Language Support

Currently supports:

• TypeScript/JavaScript (.ts, .tsx, .js, .jsx)
• Python (.py)
• C/C++ (.c, .h)

More languages coming soon!

Key Commands

Configuration

Configure Arc Flow through VS Code settings:

File → Preferences → Settings → Extensions → Arc Flow

Or edit settings.json:

{
  "arcFlow.groqApiKey": "your-groq-api-key",
  "arcFlow.geminiApiKey": "your-gemini-api-key",
  "arcFlow.aiProvider": "gemini",
  "arcFlow.awsRegion": "us-east-1",
  "arcFlow.useLSP": false
}

AI Configuration Options

• arcFlow.groqApiKey - Groq API key for fast analysis
• arcFlow.geminiApiKey - Google Gemini API key
• arcFlow.aiProvider - Choose between "gemini" or "bedrock"
• arcFlow.awsRegion - AWS region for Bedrock (default: us-east-1)
• arcFlow.bedrockModelId - Bedrock model ID
• arcFlow.awsAccessKeyId - AWS access key
• arcFlow.awsSecretAccessKey - AWS secret key
• arcFlow.useLSP - Enable LSP-based symbol resolution (experimental)

How It Works

Arc Flow uses a sophisticated multi-threaded architecture:

1. Tree-sitter Parser - Fast, incremental code parsing
2. Symbol Extraction - Identifies functions, classes, variables, imports
3. Dependency Graph - Builds relationships between code elements
4. SQLite Database - Stores symbols and relationships locally
5. AI Analysis - Provides intelligent insights on-demand
6. Interactive Visualization - React-based graph rendering

All processing happens locally on your machine. Code is never uploaded to the cloud (only minimal context is sent to AI providers when you explicitly request AI features).

Performance

• Indexing Speed: <2 minutes for 5,000 files
• Incremental Updates: <100ms for single file changes
• Symbol Queries: <10ms response time
• Memory Usage: <1GB during indexing, <500MB idle
• Graph Rendering: 60 FPS for 1,000+ nodes

Use Cases

• Onboarding: New developers understand the codebase 10x faster
• Refactoring: Assess impact before making changes
• Code Review: Identify architectural issues and technical debt
• Documentation: Auto-generate architectural insights
• Debugging: Trace function calls and dependencies

Privacy & Security

• ✅ All indexing happens locally on your machine
• ✅ No code is uploaded to the cloud
• ✅ AI features only send minimal context (the specific symbol you're analyzing)
• ✅ You control which AI provider to use
• ✅ API keys stored securely in VS Code settings
• ✅ Can be used completely offline (without AI features)

FAQ

Q: Do I need an AI API key?
A: No, core features (indexing, visualization, search) work without AI. AI features are optional enhancements.

Q: Which AI provider should I use?
A: Groq is great for fast responses (free tier available). Gemini offers the best balance of cost and quality. Bedrock is for enterprise users.

Q: Is my code safe?
A: Yes. All indexing is local. AI features only send the specific code snippet you're asking about, never your entire codebase.

Q: Can I use this offline?
A: Yes! All core features work offline. Only AI analysis requires internet.

Q: How do I handle large codebases?
A: Arc Flow is optimized for large projects. Enable file watcher for incremental updates, use Architecture view for high-level overview, and filter by directory.

Roadmap

v0.4.0

• Multi-workspace support
• Historical analysis (track code evolution over time)
• Custom health metrics and thresholds
• Export graphs to PNG/SVG

v0.5.0

• Java and Go language support
• LSP integration for accurate type resolution
• Git blame integration
• Team collaboration features

Future

• CI/CD integration
• Custom plugin system
• Cloud sync (optional)

Support

For questions, issues, or feature requests, please contact the developer or use the VS Code extension feedback system.

License

MIT License