Turn any codebase into a system design document — in minutes, inside VS Code.

Local static analysis · Interactive dependency graphs · RAG architecture chat · AI reviews

Quick Start · Features · Architecture · Case Study · Benchmarks · Changelog · Contributing

ArchLens is a production-grade VS Code extension that analyzes your repository locally, builds a structural knowledge graph, and surfaces architecture diagrams, dependency graphs, technical debt, and grounded AI chat — without uploading source code to a cloud analyzer.

Record a live demo GIF: README hero GIF guide

Architecture Dashboard	Dependency Graph	Architecture Chat
Full analysis dashboard with stats, tabs, and Mermaid diagrams	Force-directed graph — zoom, search, open files in VS Code	Grounded Q&A with confidence and file references

Architecture Diagram	Technical Debt Report	Database ERD
Pattern detection with layer flow visualization	Severity-ranked debt issues and hotspot analysis	Auto-generated entity-relationship diagram

Capture production PNGs/GIFs: docs/demo/recording-guide.md · docs/demo/screenshots/

Ask natural-language questions about your codebase. ArchLens retrieves relevant entities from a knowledge graph built during analysis, then generates answers with referenced files, dependency paths, and confidence scores.

You: Where does authentication happen?
ArchLens: JWT validation occurs in auth.middleware.ts → AuthService.verifyToken()
          Referenced: src/middleware/auth.ts, src/services/auth.service.ts
          Confidence: 87%

→ RAG pipeline docs · Demo script

A directed force graph built from import relationships and layer inference. Explore coupling visually, search nodes, and jump to source files.

→ Dependency engine docs

Deterministic detectors surface structural risks before they become production incidents:

• God classes and oversized modules
• Circular dependencies
• Layer violations (e.g. controller → repository)
• Dead code signals
• Git churn hotspots

Identifies JWT, Passport, session middleware, and login endpoints. Produces Mermaid sequence diagrams of the auth lifecycle.

Parses Prisma, TypeORM, Sequelize, and Mongoose schemas. Generates Mermaid ER diagrams from your actual model definitions.

When configured with an API key, ArchLens produces structured reviews: strengths, weaknesses, scalability risks, and actionable recommendations — grounded in deterministic analysis, not raw file dumps.

ORM and framework detection covers Express, NestJS, Next.js, React, Fastify, Prisma, TypeORM, Sequelize, Mongoose, Passport, and more.

• Node.js 18+
• npm 9+
• VS Code 1.85+

git clone https://github.com/archlens/archlens.git
cd archlens
npm install
npm run build:extension

Package and install:

cd apps/vscode-extension
npm run package
# Extensions → Install from VSIX

1. Open any repository in VS Code (monorepo root recommended for development)
2. Click the ArchLens icon in the Activity Bar
3. Click Analyze Repository
4. Explore the Dashboard — Architecture, Dependencies, Database, Auth, Debt, Review
5. Open Architecture Chat for grounded Q&A
6. Configure AI (optional): ArchLens: Configure API Key

{
  "archlens.aiProvider": "openai",
  "archlens.apiKey": "sk-...",
  "archlens.model": "gpt-4o-mini",
  "archlens.maxFiles": 10000,
  "archlens.useCache": true
}

ArchLens is a TypeScript monorepo with strict package boundaries:

archlens/
├── apps/
│   ├── vscode-extension/    # Extension host, commands, webviews
│   └── webview-ui/          # React dashboard (bundled to media/)
└── packages/
    ├── shared/              # Types, constants, shared contracts
    ├── analysis-engine/     # Scanner, AST parsers, detectors
    ├── diagram-engine/      # Dependency graphs, Mermaid generation
    ├── ai-engine/           # LLM providers, docs, export
    └── rag-engine/          # Knowledge graph, embeddings, chat

flowchart TB
  subgraph VSCode["VS Code Extension"]
    CMD[8 Commands]
    SB[Sidebar]
    DB[React Dashboard]
    CH[Architecture Chat]
  end

  subgraph Engine["Analysis Engine"]
    SCAN[Repository Scanner]
    AST[AST Parsers]
    DET[Pattern Detectors]
    DEBT[Debt Scanner]
  end

  subgraph Intelligence["AI Layer"]
    KG[Knowledge Graph]
    RAG[RAG Retriever]
    LLM[OpenAI / Claude / Gemini]
  end

  CMD --> SCAN --> AST --> DET
  DET --> DEBT
  DET --> DB
  DET --> KG --> RAG --> LLM
  CH --> RAG

Deep dives: System Architecture · Analysis Pipeline · RAG Pipeline

npm run benchmark

Full methodology: docs/benchmarks.md

Sample analyses for popular stacks:

Use these when describing ArchLens on a resume or portfolio:

• Built a VS Code extension with React webviews that analyzes codebases locally and generates architecture documentation, dependency graphs, and technical debt reports
• Designed a TypeScript monorepo with 5 domain packages (analysis, diagram, AI, RAG, shared) and 64 automated tests
• Implemented AST-based static analysis (ts-morph, Babel) with architecture pattern detection across 6+ patterns
• Built an interactive D3.js force-directed dependency graph with search, zoom, and file navigation
• Architected a RAG pipeline with knowledge graph construction, TF-IDF embeddings, and multi-provider LLM integration (OpenAI, Claude, Gemini)
• Achieved sub-second analysis on medium repos (500 files) with hash-based incremental caching

Full ATS-friendly versions: docs/resume-snippets.md

npm install
npm run build:extension   # Build all packages + extension
npm test                  # 64 unit/integration tests
npm run test:coverage     # Coverage report
npm run benchmark         # Performance benchmarks

Open the monorepo root → Run Extension → F5. See docs/contributing.md.

We welcome contributions. See CONTRIBUTING.md and docs/contributing.md.

1. Fork the repository
2. Create a feature branch
3. Run npm test and npm run build:extension
4. Open a pull request

MIT © ArchLens Contributors