pruny

Find and remove unused code in Next.js and NestJS projects.

Pruny scans your codebase using regex-based static analysis to detect unused API routes, page links, exports, files, public assets, and NestJS service methods. Works with monorepos out of the box.

Install

npm install -g pruny
# or
npx pruny

What It Detects

| Scanner | What it finds | | :------ | :------------ | | API Routes | Unused Next.js route.ts handlers and NestJS controller methods | | Broken Links | <Link>, router.push(), redirect(), href: "/path" in arrays/objects pointing to pages that don't exist | | Unused Exports | Named exports and class methods not imported anywhere | | Unused Files | Source files not reachable from any entry point | | Unused Services | NestJS service methods never called by controllers or other services | | Public Assets | Images/files in public/ not referenced in code | | Source Assets | Media files in src/ not referenced in code | | Missing Assets | References to files in public/ that don't exist |

CLI Commands

| Command | Description | | :------ | :---------- | | pruny | Interactive scan (auto-detects monorepo apps) | | pruny --all | CI mode: scan all apps, exit 1 if issues found | | pruny --fix | Interactively delete unused items | | pruny --dry-run | Simulate fix mode and output a JSON report | | pruny --app <name> | Scan a specific app in a monorepo | | pruny --folder <path> | Scan a specific folder for routes/controllers | | pruny --cleanup <items> | Quick cleanup: routes, exports, public, files | | pruny --filter <pattern> | Filter results by path or app name | | pruny --ignore-apps <list> | Skip specific apps (comma-separated) | | pruny --no-public | Skip public asset scanning | | pruny --json | Output results as JSON | | pruny -v, --verbose | Verbose debug logging | | pruny --dir <path> | Set target directory (default: ./) | | pruny -c, --config <path> | Path to config file | | pruny init | Generate a pruny.config.json with defaults |

Configuration

Create pruny.config.json in your project root (or run pruny init):

{
  "ignore": {
    "routes": ["/api/webhooks/**", "/api/cron/**"],
    "folders": ["node_modules", ".next", "dist"],
    "files": ["*.test.ts", "*.spec.ts"],
    "links": ["/custom-path", "/legacy/*"]
  },
  "extensions": [".ts", ".tsx", ".js", ".jsx"]
}

Ignore Options

| Key | What it does | Example | | :-- | :----------- | :------ | | ignore.routes | Skip API routes matching these patterns | ["/api/webhooks/**"] | | ignore.folders | Exclude directories from scanning | ["node_modules", "dist"] | | ignore.files | Exclude specific files from scanning | ["*.test.ts"] | | ignore.links | Suppress broken link warnings for these paths | ["/view_seat", "/admin/*"] |

All patterns support glob syntax (* matches any characters, ** matches nested paths).

Pruny also reads .gitignore and automatically excludes those folders.

Additional Config Options

| Key | What it does | | :-- | :----------- | | nestGlobalPrefix | NestJS global route prefix (e.g., "api/v1") | | extraRoutePatterns | Additional glob patterns to detect route files | | excludePublic | Set true to skip public asset scanning |

Config File Locations

Pruny searches for config files recursively across your project:

• pruny.config.json
• .prunyrc.json
• .prunyrc

In monorepos, configs from multiple apps are merged together. CLI --config takes precedence.

Multi-Tenant / Subdomain Routing

Pruny automatically handles multi-tenant architectures where routes live under dynamic segments like [domain].

For example, if your file structure is:

app/(code)/tenant_sites/[domain]/view_seat/page.tsx

And your components reference /view_seat (resolved at runtime via subdomain), Pruny recognizes this as a valid route and will not report it as a broken link. The matched tail must contain at least one literal segment (e.g., view_seat) — fully-dynamic tails like [token] alone won't match arbitrary paths.

If auto-detection doesn't cover your case, use ignore.links in config:

{
  "ignore": {
    "links": ["/view_seat", "/review", "/custom-path"]
  }
}

Monorepo Support

Pruny auto-detects monorepos by looking for an apps/ directory. It scans each app independently but checks references across the entire monorepo root.

# Scan all apps (CI-friendly, exits 1 on issues)
pruny --all

# Scan a specific app
pruny --app web

# Skip certain apps
pruny --ignore-apps admin,docs

CI Integration

Add to your CI pipeline to catch unused code before merging:

npx pruny --all

This scans all monorepo apps and exits with code 1 if any issues are found. Combine with --json for machine-readable output.

How It Works

1. Route Detection: Finds all app/api/**/route.ts (Next.js) and *.controller.ts (NestJS) files
2. Link Detection: Finds <Link>, router.push(), redirect(), href: "/path" in arrays/config objects, <a> tags, revalidatePath(), and pathname === comparisons — validates all against known page routes. The summary table always shows an "Internal Links" row when links are scanned, so you can see the feature is active
3. Reference Scanning: Searches the entire codebase for string references to routes, exports, and assets
4. Dynamic Route Matching: Understands [id], [...slug], [[...slug]] dynamic segments
5. Fix Mode: Removes unused methods, exports, and files with a cascading second pass to catch newly dead code

Vercel Cron Detection

Routes listed in vercel.json cron jobs are automatically marked as used:

{ "crons": [{ "path": "/api/cron/cleanup", "schedule": "0 0 * * *" }] }

Debug Mode

DEBUG_PRUNY=1 pruny

Enables verbose logging across all scanners.

License

MIT