██████╗██╗██████╗ ██╗    ██╗███████╗
  ██╔════╝██║██╔══██╗██║    ██║██╔════╝
  ██║     ██║██████╔╝██║ █╗ ██║█████╗  
  ██║     ██║██╔═══╝ ██║███╗██║██╔══╝  
  ╚██████╗██║██║     ╚███╔███╔╝███████╗
   ╚═════╝╚═╝╚═╝      ╚══╝╚══╝ ╚══════╝

The Web Vitals for the AI Web

Make your website visible and answerable to AI agents, LLMs, and answer engines.

Website · Report Bug · Request Feature

What is CIPWE?

CIPWE is the ESLint for AI-readability. It scans your website and tells you how well AI agents (ChatGPT, Perplexity, Claude, Gemini) can understand, crawl, and answer questions about your content.

Just like Lighthouse measures web performance and Web Vitals measure user experience — CIPWE measures AI-readiness.

npx cipwe audit https://your-site.com

  CIPWE Score    72/100  (72%)
  Grade          🔵 B

  ✔ Structured Data       █████████████░░░░░░░ 20/30
  ✔ Semantic HTML          ███████████████░░░░░ 15/20
  ✔ Metadata               ████████████████████ 15/15
  ◐ Crawl Signals          ██████████████░░░░░░ 10/15
  ◐ Content Clarity        ██████████████░░░░░░  7/10
  ✘ Agent Signals          ██████████░░░░░░░░░░  5/10

Why CIPWE?

The web is shifting. AI agents are the new search engines:

• ChatGPT browses the web to answer questions
• Perplexity indexes websites for instant answers
• Claude reads pages for research and analysis
• Gemini summarizes web content in search results

If your site isn't AI-readable, you're invisible to the next generation of search.

CIPWE helps you fix that in 60 seconds.

Quick Start

# Audit any website
npx cipwe audit https://example.com

# Audit local project
npx cipwe audit .

# Auto-generate fixes
npx cipwe fix https://example.com

# Quick score check
npx cipwe score https://example.com

# JSON output for CI/CD
npx cipwe audit https://example.com --json

Installation

# Use directly with npx (no install needed)
npx cipwe audit https://your-site.com

# Or install globally
npm install -g cipwe

# Or add to project
npm install --save-dev cipwe

The CIPWE Score

CIPWE evaluates your site across 6 categories with 20 rules, totaling 100 points:

| Category | Weight | What It Measures | |----------|--------|-----------------| | Structured Data | 30 pts | JSON-LD, Schema.org types, FAQ/Article schemas | | Semantic HTML | 20 pts | H1 tags, heading hierarchy, <main>, <article> | | Metadata | 15 pts | Title, meta description, OpenGraph tags | | Crawl Signals | 15 pts | robots.txt, sitemap.xml, canonical URLs | | Content Clarity | 10 pts | Content density, lists/tables, Q&A structure | | Agent Signals | 10 pts | llms.txt, machine-readable summaries |

Grading Scale

| Grade | Score | Meaning | |-------|-------|---------| | 🟢 A+ | 95-100 | AI-ready. Top-tier. | | 🟢 A | 90-94 | Excellent AI visibility | | 🔵 B+ | 80-89 | Good, minor improvements needed | | 🔵 B | 70-79 | Decent, several improvements possible | | 🟡 C | 60-69 | Moderate, significant gaps | | 🟠 D | 40-59 | Poor AI-readability | | 🔴 F | 0-39 | Not AI-readable |

Commands

cipwe audit <target>

Scan a website and generate a full CIPWE report.

cipwe audit https://example.com        # Deployed site
cipwe audit example.com                # Auto-adds https://
cipwe audit .                          # Local HTML files
cipwe audit ./dist                     # Build output
cipwe audit https://example.com --json # JSON for CI/CD

cipwe fix <target>

Auto-generate missing files to improve your CIPWE score.

cipwe fix https://example.com
cipwe fix . -o ./public

Generates:

• llms.txt — AI agent directive file
• structured-data.jsonld — JSON-LD schema
• structured-data-snippet.html — Copy-paste HTML snippet
• robots.txt — AI-bot-friendly robots file
• sitemap.xml — Auto-generated sitemap
• cipwe-report.md — Detailed audit report

cipwe score <target>

Quick one-line score check.

$ cipwe score https://example.com
🔵 CIPWE 72/100 (B)

CI/CD Integration

GitHub Actions

name: CIPWE Audit
on: [push]

jobs:
  cipwe:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - uses: actions/setup-node@v4
        with:
          node-version: '20'
      - run: npx cipwe audit https://your-site.com --json > cipwe-report.json
      - name: Check CIPWE Score
        run: |
          SCORE=$(cat cipwe-report.json | jq '.percentage')
          echo "CIPWE Score: $SCORE%"
          if [ "$SCORE" -lt 50 ]; then
            echo "CIPWE Score below threshold!"
            exit 1
          fi

What is llms.txt?

llms.txt is an emerging standard (like robots.txt but for AI). It tells AI agents:

• What your site is about
• What content is available
• How AI should use your content

CIPWE auto-generates this for you with the fix command.

Rules Reference

| Rule | Points | Description | |------|--------|-------------| | JSON-LD Presence | 10 | Has <script type="application/ld+json"> | | JSON-LD Valid Type | 8 | Uses recognized Schema.org @type | | FAQ Schema | 6 | FAQPage schema for Q&A content | | Article/Product Schema | 6 | Rich content type schemas |

| Rule | Points | Description | |------|--------|-------------| | Single H1 | 5 | Exactly one <h1> tag | | Heading Hierarchy | 5 | Proper H1→H2→H3 nesting | | Has <main> | 5 | Uses semantic <main> element | | Has <article>/<section> | 5 | Semantic sectioning elements |

| Rule | Points | Description | |------|--------|-------------| | Page Title | 5 | Meaningful <title> tag (50-60 chars) | | Meta Description | 5 | Descriptive meta description (150-160 chars) | | OpenGraph Tags | 5 | og:title, og:description, og:image |

| Rule | Points | Description | |------|--------|-------------| | robots.txt | 5 | Accessible with AI bot directives | | sitemap.xml | 5 | Valid sitemap with URLs | | Canonical URL | 5 | <link rel="canonical"> present |

| Rule | Points | Description | |------|--------|-------------| | Content Length | 3 | 300+ words of readable content | | Lists & Tables | 3 | Uses <ul>, <ol>, <table> | | Q&A Structure | 4 | Question headings, FAQ patterns |

| Rule | Points | Description | |------|--------|-------------| | llms.txt | 5 | AI directive file at site root | | Structured Summary | 5 | Rich meta signals (lang, author, etc.) |

Philosophy

SEO optimized for Google.
CIPWE optimizes for the AI web.

The web is becoming answer-first. Users don't click links — they ask AI agents. Your content needs to be:

1. Findable — AI agents can discover it
2. Parseable — Structured for machine reading
3. Answerable — Clear enough to generate accurate answers

CIPWE measures all three.

Contributing

Contributions welcome! See CONTRIBUTING.md for guidelines.

git clone https://github.com/user/cipwe.git
cd cipwe
npm install
npm run build
node dist/index.js audit https://example.com

License

MIT © CIPWE

Make your site answerable.

cipwe.someshghosh.me