@nemo-cli/deps

Dependency analysis CLI for Next.js App Router projects

Install

$ pnpm add @nemo-cli/deps --global

Usage

# Analyze a Next.js project
nd analyze <path>

# Analyze src folder (still writes ai-docs under project root)
nd analyze ./src

# Query page component tree
nd page <route>

# Output as Graphviz DOT
nd analyze <path> --format dot --output deps.dot

# Output as JSON
nd analyze <path> --format json

# Output AI-friendly JSON (page map + node map)
nd analyze <path> --format ai --output ./ai-docs/deps.ai.json

# Run AI analysis with progress UI and per-component files
nd ai --input ./ai-docs/deps.ai.json

# Analyze specific route
nd analyze <path> --route /dashboard

# Show only leaf nodes
nd analyze <path> --leaves

# Show only orphan nodes
nd analyze <path> --orphans

# Highlight cycles in output
nd analyze <path> --cycles

# Limit analysis depth
nd analyze <path> --max-depth 3

# Follow external dependencies
nd analyze <path> --external

# Verbose output
nd analyze <path> --verbose

Features

• Next.js App Router Support: Automatically detects and analyzes Next.js App Router routes
• Multiple Output Formats: Tree (default), JSON, Graphviz DOT
• AI Output: Page map + node map + cycle groups (tree output stays in --format tree)
• AI Analysis Runner: Interactive route selector + progress view + per-component outputs
• Cycle Detection: Identifies circular dependencies
• Leaf/Orphan Detection: Finds unused or terminal dependencies
• Configurable Depth: Limit analysis to specific depth levels
• Module System Detection: Supports ES6 modules, CommonJS, and dynamic imports

Output Formats

Tree

Hierarchical tree visualization with emoji markers for different node types:

• 🚀 Entry points (routes/pages)
• 🍃 Leaf nodes (no dependencies)
• ⚠️ Dynamic imports
• Dashed lines: CommonJS modules

JSON

Structured JSON output with full graph data, statistics, and route metadata.

DOT

Graphviz DOT format for generating visual dependency graphs:

nd analyze . --format dot --output deps.dot
dot -Tsvg deps.dot -o deps.svg

Commands

analyze [options] <path>

Analyzes dependencies of a Next.js project.

Options:

• --format <type> - Output format: tree, json, dot, or ai (default: tree)
• --output <file> - Output file path
• --route <path> - Analyze specific route only
• --leaves - Show only leaf nodes (no dependencies)
• --orphans - Show only orphan nodes (no importers)
• --cycles - Highlight circular dependencies
• --max-depth <number> - Limit analysis depth
• --external - Follow external dependencies
• --no-ai-json - Disable default ai-docs/deps.ai.json side output
• --verbose - Enable verbose output

Default behavior:

• nd analyze always writes ai-docs/deps.ai.json under project root
• nd analyze --format ai writes main output to ai-docs/deps.ai.json by default

page [options] <route>

Queries page component tree from generated JSON.

Options:

• --input <file> - Input JSON file path (default: deps.json)
• --format <type> - Output format: tree or json (default: tree)

Examples

# Basic analysis
nd analyze ./my-nextjs-app

# Generate SVG visualization
nd analyze ./my-nextjs-app --format dot --output deps.dot
dot -Tsvg deps.dot -o deps.svg

# Analyze specific route
nd analyze ./my-nextjs-app --route /dashboard

# Find circular dependencies
nd analyze ./my-nextjs-app --cycles

# Quick check for leaf nodes
nd analyze ./my-nextjs-app --leaves --format tree

# Query page component tree
nd page /dashboard --input deps.json

License

ISC © gaozimeng

ai [options]

Runs AI analysis for a selected page route and stores per-component results.

Options:

• --input <file> - Input AI JSON file (default: ai-docs/deps.ai.json)
• --output <dir> - Output directory (default: /ai-docs)
• --route <path> - Analyze a specific route (skip selector)
• --engine <engine> - AI engine: openai, google, zhipu, none
• --model <name> - Model name
• --max-source-chars <number> - Max source chars per component (default: 12000)
• --fresh - Ignore existing analysis results

Output:

• Component files: ai-docs/components/**
• Page index: ai-docs/pages/<route>/page.json

AI Keys:

• OPENAI_API_KEY for OpenAI
• GOOGLE_API_KEY for Google Gemini
• ZHIPU_API_KEY for Zhipu (glm) If no --engine is provided, the runner will auto-detect based on available API keys. Optional:
• ZHIPU_BASE_URL for custom Zhipu endpoint