GEO Toolkit

By Gábor Krasznai 👁 5 views ▲ 1 votes

GEO toolkit for AI search optimization. Audit content for AI citation readiness (ChatGPT Search, Perplexity, Google AI Overviews, Claude, Copilot) and generate schema.org JSON-LD markup for Article, FAQPage, Organization, Person, HowTo, and Service. Two tools, six schema types, Hungarian and English support. Grounded in 2026 AI-search benchmarks and peer-reviewed GEO research. Runs fully offline — no API keys, no external calls.

openclaw plugins install --path ./openclaw-geo-toolkit

{
  "plugins": {
    "allow": ["geo-toolkit"],
    "entries": {
      "geo-toolkit": {
        "enabled": true
      }
    }
  }
}

# GEO Toolkit — OpenClaw Plugin

Two tools for **Generative Engine Optimization (GEO)** — audit content for AI search citation readiness and generate ready-to-paste schema.org JSON-LD markup. Built and maintained by [SocialPro](https://www.socialpro.hu), a Hungarian AI marketing and GEO specialist agency.

> *Magyar olvasóknak: [README magyarul →](./README.hu.md)*

[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](./LICENSE)
![OpenClaw Plugin](https://img.shields.io/badge/OpenClaw-Plugin-orange)
![TypeScript](https://img.shields.io/badge/TypeScript-strict-blue)

## Why GEO matters

Generative Engine Optimization is the practice of structuring content so it gets **cited** by AI search engines (ChatGPT Search, Perplexity, Google AI Overviews, Claude, Microsoft Copilot) — not just ranked in traditional search. The 2026 data shows how critical this has become:

- **AI search traffic grew 527%** year-over-year (Previsible, 2025)
- **AI Overviews cover 58% more queries** year-over-year — healthcare 88%, B2B tech 82% (BrightEdge, 2026)
- **AI-search visitors convert 4.4× better** than traditional organic traffic (Semrush, 2025)
- **69% of searches are zero-click** (Similarweb, 2025)
- **Adding statistics improves AI citation rate by +34%** (peer-reviewed GEO research, 2024-2025)

For full-service GEO audits, content strategy, and implementation, see our [AI marketing and GEO consulting service](https://www.socialpro.hu/szolgaltatasok/ai-marketing).

## Installation

Install from a local clone (recommended for development):

```bash
git clone https://github.com/socialproKGCMG/openclaw-geo-toolkit.git
openclaw plugins install --path ./openclaw-geo-toolkit
```

Enable the plugin in your `openclaw.json`:

```json
{
  "plugins": {
    "allow": ["geo-toolkit"],
    "entries": {
      "geo-toolkit": {
        "enabled": true
      }
    }
  }
}
```

No configuration is required — the plugin runs fully offline with no API keys or external calls.

## Tool 1: `geo_audit`

Audit any text or HTML content against a 6-dimension GEO rubric and get a 0-100 score plus actionable recommendations.

### Signature

```
geo_audit({ content: string, format?: "text" | "html" })
```

| Parameter | Type   | Default  | Description                                                        |
|-----------|--------|----------|--------------------------------------------------------------------|
| `content` | string | —        | The content to audit. Plain text or HTML.                          |
| `format`  | string | `"text"` | Use `"html"` when your input has tags (enables full scoring).      |

### Scoring dimensions

| Dimension             | Weight | What it measures                                                 |
|-----------------------|-------:|------------------------------------------------------------------|
| Citation Readiness    |   25%  | Fact density, 134-167 word citable blocks, opening hook          |
| Answer Alignment      |   20%  | First 100 words vs. H1 keyword overlap, Quick Answer presence   |
| Content Authority     |   20%  | Outbound links, source citations, E-E-A-T markers               |
| Entity Integration    |   15%  | JSON-LD schema markup presence and types                         |
| Technical AI Access   |   10%  | Heading hierarchy validity, H1 uniqueness                        |
| Competitive Edge      |   10%  | Unique value signals, buzzword avoidance, passive voice penalty |

### Score tiers

- **85-100 — AI-Optimized**: content is ready for AI citation
- **70-84 — AI-Ready**: minor tuning recommended
- **55-69 — Needs Improvement**: structural and content work required
- **< 55 — Not Optimized**: major rewrite needed

### Example output

```json
{
  "overallScore": 72,
  "category": "AI-Ready",
  "dimensions": [
    {
      "key": "citability",
      "label": "Citation Readiness",
      "weight": 0.25,
      "score": 78,
      "issues": [],
      "suggestions": [
        "Increase fact density to 5-7 verifiable facts per 1000 words."
      ]
    }
  ],
  "recommendations": [
    "[Entity Integration] Add Organization schema for entity identification.",
    "[Content Authority] Cite named sources with the format 'Author, Year'.",
    "[Citation Readiness] Increase fact density to 5-7 verifiable facts per 1000 words."
  ],
  "meta": {
    "wordCount": 1450,
    "format": "html",
    "h1Count": 1,
    "factDensityPer1000Words": 4.1
  }
}
```

Schema markup is one of the foundations of technical GEO — [see our blog for deeper articles on GEO implementation](https://www.socialpro.hu/blog).

## Tool 2: `geo_generate_schema`

Generate a ready-to-paste JSON-LD schema.org markup block for the six most citation-relevant types.

### Signature

```
geo_generate_schema({ type: SchemaType, data: object, wrap?: boolean })
```

| Parameter | Type    | Default | Description                                                        |
|-----------|---------|---------|--------------------------------------------------------------------|
| `type`    | string  | —       | One of `Article`, `FAQPage`, `Organization`, `Person`, `HowTo`, `Service`. |
| `data`    | object  | —       | Fields for the schema. See examples below.                         |
| `wrap`    | boolean | `true`  | Wrap output in `<script type="application/ld+json">`.              |

### Example — FAQPage (highest AI citation rate: 58%)

```json
{
  "type": "FAQPage",
  "data": {
    "questions": [
      { "q": "What is GEO?", "a": "Generative Engine Optimization is the practice of optimizing content for AI search engines." },
      { "q": "How is GEO different from SEO?", "a": "SEO targets search rankings; GEO targets citation inside AI-generated answers." }
    ]
  }
}
```

### Example — Article

```json
{
  "type": "Article",
  "data": {
    "headline": "The Complete Guide to GEO in 2026",
    "description": "Learn how to optimize content for AI search engines.",
    "author": { "name": "Gábor Krasznai", "url": "https://www.socialpro.hu" },
    "datePublished": "2026-01-15",
    "image": "https://example.com/cover.jpg",
    "publisher": { "name": "SocialPro", "logo": "https://www.socialpro.hu/logo.png" }
  }
}
```

The scoring methodology and schema templates are documented in depth across [SocialPro's AI visibility case studies](https://www.socialpro.hu/esettanulmanyok).

## Use cases

- **Pre-publish check** — run `geo_audit` on a draft before publishing to catch weak opening paragraphs, missing schema, or low fact density.
- **Competitor analysis** — audit competing articles to see where they score well and where yours can differentiate.
- **Schema generation for CMS** — generate JSON-LD markup in your AI agent workflow and paste it into WordPress, Astro, Next.js, or any CMS.
- **Hungarian content optimization** — both English and Hungarian content are supported; the fact-density and keyword-overlap heuristics work in both languages.
- **Bulk content audits** — combine with an agent that fetches multiple URLs and runs the audit across a content library.

## Methodology

The 6-dimension rubric is derived from multiple sources:

- **BrightEdge 2026** — AI Overview citation patterns, multimodal content weight
- **Semrush AI search analysis** (2025) — conversion data, citation rates per platform
- **Previsible 2025** — YoY AI search traffic growth, platform overlap analysis
- **Peer-reviewed GEO research** (Aggarwal et al., KDD 2024) — Position-Adjusted Word Count and Subjective Impression metrics form the academic foundation
- **Internal GEO research at SocialPro** — Hungarian-market case studies

Scoring is rule-based, deterministic, and runs fully offline. Same input always produces the same output — ideal for CI pipelines and regression testing of content.

## About the author

This plugin is built and maintained by [SocialPro / KG Creative Media Group Kft.](https://www.socialpro.hu/rolunk) — a Hungarian AI automation and marketing agency specialising in Generative Engine Optimization, SEO, Meta advertising, and custom AI integrations for Hungarian SMEs.

## Support and custom audits

Bug reports and feature requests: [GitHub Issues](https://github.com/socialproKGCMG/openclaw-geo-toolkit/issues).

For custom GEO audits, enterprise deployments, or training for your team, [contact SocialPro](https://www.socialpro.hu/kapcsolat).

## License

MIT License — see [LICENSE](./LICENSE) for details.

Copyright © 2026 KG Creative Media Group Kft. (SocialPro)