geo-audit
Evaluates any website's AI visibility with 15 checks across crawlability, structure, content, and connectivity, and provides actionable fixes.
README
geo-audit
Free GEO (Generative Engine Optimization) audit for any website. No API keys, no login, no tracking.
AI search engines — ChatGPT, Claude, Perplexity, Gemini — are how people increasingly find businesses. If those engines can't reach your site, can't tell who you are, or can't quote you, you don't exist in their answers. Big companies pay enterprise tools to fix this. This tool gives the same checklist to everyone.
Built by Enso Labs for community builders, small businesses, and anyone who wants to own their own AI visibility.
$ npx tsx src/cli.ts yoursite.org
GEO AUDIT https://yoursite.org
Score 71/100 — C — visible but under-represented in AI answers
✅ Site loads over HTTPS 5/5
🟡 AI crawlers allowed in robots.txt 8/15
🟡 llms.txt for AI assistants 4/8
🟡 JSON-LD structured data (schema.org) 8/15
...
Top fix: Explicitly allow the AI engines you want to be represented in.
What it checks (15 checks, 4 pillars)
| Pillar | Checks |
|---|---|
| 1 · Crawlability — can AI engines reach you? | HTTPS, robots.txt, AI-crawler allowlist (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, +6 more), sitemap, llms.txt |
| 2 · Structure — can they understand who you are? | JSON-LD identity (Organization/Person), canonical URL, Open Graph, title/meta, indexability |
| 3 · Content — is your writing quotable? | Single H1, definition-lead sentence ("X is a …" — the shape AI answers quote), FAQ schema |
| 4 · Connectivity — feeds & agent access | RSS/Atom feed, MCP manifest (/.well-known/mcp.json) |
Every failed check comes with a plain-English fix and a copy-paste snippet (a ready robots.txt block, an llms.txt template, a JSON-LD starter).
Audit in the browser (no install)
Go to geo.ensolabs.ai, type your URL, press AUDIT ▶. You get:
- the full 15-check report with copy-paste fixes
- a per-engine breakdown — ChatGPT, Claude, Perplexity, and Gemini each scored separately (each engine reads your site with a different crawler, so the answers genuinely differ)
- a live "how does AI describe you today" check (Perplexity, rate-limited)
- Notion mode — audit any
*.notion.siteand see which fixes you control inside Notion vs. which need a custom domain
The tools page has free no-code generators for everything the audit asks for: a robots.txt AI-allowlist builder, an llms.txt generator, identity JSON-LD, and FAQ schema.
Quick start
git clone https://github.com/nycsav/notion-geo-audit && cd notion-geo-audit
npm install
npx tsx src/cli.ts yoursite.org # terminal scorecard
npx tsx src/cli.ts yoursite.org --md report.md # full report with fixes
npx tsx src/cli.ts yoursite.org --json # for pipelines/agents
Use it inside Claude (the GEO connector)
geo-audit ships as an MCP server, so your community can plug it straight into Claude and just ask:
"Audit mysite.org for AI visibility and walk me through the top 3 fixes."
npm run build
claude mcp add geo-audit -- node /path/to/notion-geo-audit/dist/mcp.js
Tools exposed:
audit_site— the full scored audit as markdowncheck_ai_representation— asks Perplexity, live, "how do AI engines describe this site today, and which sources do they trust?" (optional — setPERPLEXITY_API_KEY)
The second one matters: it shows you the gap between who you are and how AI describes you — including when engines confuse you with a similarly-named entity.
Optional: live representation check
export PERPLEXITY_API_KEY=pplx-...
npx tsx src/cli.ts yoursite.org --representation
Why we built this
Enso Labs took its own site from invisible to 100/100 on these checks — schema on every page, llms.txt, AI-crawler policy, definition-lead copy, an MCP endpoint. The method was the easy part; it just lived in private audits. This tool open-sources the method so any community — a forum, a small business group, a local org — can run it on their own sites and own their own representation.
Roadmap: GA4 traffic correlation (see which AI engines actually send you visitors) · Notion sync (audit history as a database) · scheduled re-audits.
License
MIT — use it, fork it, run it for your community.
Powered by Enso Labs · Built with Claude Code
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。