geoready-mcp
# 🛰️ GeoReady MCP Server (`geoready-mcp`)
**Zero-dependency Model Context Protocol (MCP) server for 2026 Generative Engine Optimization (GEO), ChatGPT Shopping Schema & AI Agent Readiness auditing.**
Let **Claude Desktop**, **Claude Code**, **Cursor**, and **Windsurf** inspect any live website across 8 deterministic GEO benchmarks and generate ready-to-deploy `robots.txt`, `/llms.txt`, `Content-Signal` HTTP headers, and `JSON-LD` structured data.
> 🌐 **Interactive Web Scanner & Multi-Page Sitemap Batch Suite**: **[https://geoready.aitruelabs.com](https://geoready.aitruelabs.com)**
---
## Quick Start (`uvx` — Zero Install)
Add to your `claude_desktop_config.json` or `.cursor/mcp.json`:
```json
{
"mcpServers": {
"geoready-auditor": {
"command": "uvx",
"args": [
"--from",
"git+https://github.com/yixuanrobot/geoready-mcp",
"geoready-mcp"
]
}
}
}
```
## What `audit_website_geo_readiness` Checks (0–100 Score)
1. **AI Crawler Access & `robots.txt`** (`OAI-SearchBot`, `GPTBot`, `ChatGPT-User`, `PerplexityBot`, `ClaudeBot`, `Google-Extended` + firewall challenge detection)
2. **`/llms.txt` & `/llms-full.txt` Specification** (detects broken SPA HTML fallbacks)
3. **`Accept: text/markdown` Negotiation & Token Bloat %**
4. **JSON-LD Knowledge Graph & ChatGPT Shopping Offers** (`Organization`, `Product`/`SoftwareApplication`, `FAQPage`)
5. **2026 `Content-Signal` HTTP Headers** (`ai-train=yes, search=yes, ai-input=yes`)
6. **Citation Metadata & Semantic Headings**
7. **Structured Table & Quantifiable Fact Density**
8. **XML Sitemap Discovery**
---
## License
MIT © [AI True Labs](https://aitruelabs.com) (`support@aitruelabs.com`)
TDQS
Scored across 1 tool
With only one tool in the set, there is no possibility of misselection or overlapping purpose. The tool's scope (a GEO/ChatGPT Shopping readiness audit on a URL) is explicit and self-contained.
The single name, audit_website_geo_readiness, follows a clear snake_case verb_noun_resource pattern and is descriptive. It cannot demonstrate consistency across a set, so it falls just short of a 5.
One tool is thin for a server, even a narrowly scoped one; the 8-point audit is bundled into a single call with no granular options. It is defensible for a single-purpose auditor but leaves no room for narrower operations.
The audit itself appears comprehensive (crawler access, llms.txt, markdown tokens, Content-Signal, JSON-LD) and returns remediation snippets, covering the core purpose. However, there is no way to re-audit after fixes, batch multiple URLs, or retrieve prior results, creating dead ends for follow-up workflows.