SEO + AdSense Compliance Inspector

A VS Code extension that statically scans HTML, PHP/CodeIgniter 4, React/JSX, Vue, Svelte, and Markdown files for SEO problems, Google AdSense policy risks, structured-data gaps, and Core Web Vitals issues — and optionally uses VS Code's Language Model API (e.g. GitHub Copilot Chat) to suggest one-click fixes, simulate an AdSense reviewer, and estimate how AI-written your prose looks.

Static analysis runs locally with zero runtime dependencies and no telemetry. AI features are entirely optional and only activate when you trigger them.

What it actually does

1. Static analyzers (always on, no AI required)

Issues appear as inline squiggles and in the Problems panel. Severity ranges from error → warning → info.

The analyzers operate on the source text using targeted regex (see Why no runtime dependencies). They run on open, save, and — if seoAdsense.enableAutoScan is on — debounced on edit.

2. Status bar indicators

When a supported file is active, three items appear on the right of the status bar:

• SEO score (0–100) with pass / warning / error icon. Click to open the report panel.
• AI-written % after you run the AI-written detection on the file (cached per file version).
• Issue counter showing errors✗ warnings⚠. Click to open the report panel.

3. AI features (require a VS Code Language Model provider)

These all go through vscode.lm — typically powered by GitHub Copilot Chat. If no model is available, the AI features quietly disable themselves and static analysis keeps working. The extension never asks for an API key.

• ✨ Fix with AI — Lightbulb action on every diagnostic. The model generates a targeted fix, which is shown for confirmation before being applied.
• Fix All Issues with AI — Walks through every diagnostic in the active file in sequence, asking for confirmation on each.
• Simulate Google AdSense Review — Opens a WebView panel with the readiness score, issue breakdown by category, and a structured LLM-generated verdict (strengths, concerns, action items, and specific policy risks). This is a heuristic simulation, not an official Google review.
• Export report as PDF — The review panel has an Export PDF button that saves a self-contained, print-styled HTML report and opens it in your browser. The browser print dialog launches automatically so you can pick Save as PDF as the destination.
• Detect AI-Written Likelihood — Sends the visible prose (with HTML/templating stripped) to the model and returns a 0–100 score plus a one-sentence reason. The result is cached and shown in the status bar.

Commands

All commands are available from the Command Palette under the SEO/AdSense category. Some also appear in the editor title bar and right-click menu on supported file types.

Supported languages: html, php, javascriptreact, typescriptreact, vue, svelte, markdown.

Configuration

Open Settings → search "SEO/AdSense":

Install

From the Marketplace

Search "SEO AdSense Compliance Inspector" (publisher MMLTECH) in the Extensions view.

From source

git clone https://github.com/mmlTools/seo-adsense-inspector.git
cd seo-adsense-inspector
npm install
npm run compile
# Open in VS Code and press F5 to launch the Extension Development Host

Package locally

npm install
npm run package   # uses @vscode/vsce to produce a .vsix
code --install-extension seo-adsense-inspector-*.vsix

Project structure

The extension is written in TypeScript and compiled to out/.

seo-adsense-inspector/
├── package.json
├── tsconfig.json
├── src/
│   ├── extension.ts                       # Activation, commands, diagnostics, status bar
│   ├── types.ts
│   ├── analyzers/
│   │   ├── index.ts                       # Orchestrates all analyzers
│   │   ├── issues.ts                      # Issue codes, severities, fix prompts
│   │   ├── score.ts                       # 0-100 readiness score
│   │   ├── seoAnalyzer.ts
│   │   ├── contentAnalyzer.ts
│   │   ├── adsenseAnalyzer.ts
│   │   ├── schemaAnalyzer.ts
│   │   └── performanceAnalyzer.ts
│   ├── llm/
│   │   ├── aiFixer.ts                     # vscode.lm: per-issue and "fix all" flows
│   │   └── aiWrittenDetector.ts           # AI-written-likelihood scoring
│   ├── providers/
│   │   └── codeActionProvider.ts          # Lightbulb code actions
│   ├── views/
│   │   └── adsenseReviewPanel.ts          # The review / report WebView
│   └── utils/
│       └── htmlUtils.ts                   # Regex-based HTML extraction
└── samples/
    └── test.html

Why no runtime dependencies

The extension ships zero runtime npm packages. HTML is parsed with targeted regex because:

1. We only need approximate locations and the visible-text content, not a real DOM.
2. Source files often contain templating noise (PHP, JSX, Liquid, Twig) that defeats real HTML parsers.
3. It keeps the bundle tiny, the supply-chain surface minimal, and activation fast.

The trade-off: pathological or unusual markup can produce false positives. Issue reports welcome.

(npm install is only required to build from source — TypeScript and @vscode/vsce are devDependencies.)

Privacy

• Static analysis runs entirely locally. Nothing leaves your machine.
• AI features send the relevant snippet (and, for the AI-written detector, up to ~6 KB of visible prose) to whichever model provider you have configured in VS Code. The extension itself collects no telemetry.

Known limitations

• The Simulate AdSense Review verdict is an LLM simulation, not an official Google decision.
• Prohibited-content detection is keyword-based and will produce false positives on legitimate writing about restricted topics.
• The extension analyzes one file at a time. Site-wide concerns (cross-page duplication, sitemap, robots.txt, server response headers) are out of scope.
• AI-written likelihood is a stylistic heuristic from an LLM — treat it as a smell test, not a verdict.

Contributing

Issues and pull requests are welcome at https://github.com/mmlTools/seo-adsense-inspector.

License

MIT