🚀 open-geo

Open-source, AI-Native Generative Engine Optimization (GEO) & Technical Audit Suite.
Like Lighthouse for Google SEO, open-geo is the open standard toolkit for Generative Engine Optimization (ChatGPT Search, Perplexity, Google AIO, Claude). Open-source AI visibility checker for the ChatGPT era: audit how AI assistants see your brand, then generate the fixes (llms.txt, robots.txt, Schema.org) that unblock them.

Website: geofn.com · Compare SaaS tools: Profound & OtterlyAI alternatives

---

<p align="center"> <img src="assets/terminal-preview.svg" alt="open-geo CLI terminal preview" width="850" /> </p> <p align="center"> _{🖥️ <em>Runs 100% locally with zero data uploaded. Prefer a visual browser playground? Check out the <a href="https://geofn.com/#demo"><b>Interactive Live Demo →</b></a></em>} </p>

---

⚡ Quick Start (No Installation Required)

Audit any website in seconds directly via npx:

# 🛠️ Audit your website's GEO readiness (robots.txt, llms.txt, Schema.org, AI crawlability)
npx open-geo audit https://yourdomain.com

# 📑 Generate standard /llms.txt and /llms-full.md specifications
npx open-geo generate-llms https://yourdomain.com --full

# ❓ Extract FAQs & generate Schema.org FAQPage JSON-LD
npx open-geo generate-faq https://yourdomain.com

# 🔑 Mine real-time buyer-intent keywords from Google & Bing
npx open-geo keywords "best crm for startups"

# 🌐 Decompose 5-dimensional Query Fan-out intents
npx open-geo fanout "ai video generator" --domain yourdomain.com

# 🔍 Probe single-model AI visibility & buyer-funnel awareness (zero-token heuristic or live LLM)
npx open-geo probe "yourbrand" -d yourdomain.com

# 🔌 Start local Model Context Protocol (MCP) Server for Cursor & Claude Code
npx open-geo mcp

⚙️ Environment Configuration (Optional)

open-geo works out-of-the-box with zero configuration for static technical audits. For live single-model LLM probing (open-geo probe), simply configure your preferred provider's API key:

cp .env.example .env
# Recommended: Set OPENROUTER_API_KEY for 100+ models with 1 key (or direct GEMINI_API_KEY / DEEPSEEK_API_KEY)

---

🌟 Core Features (Community Edition)

• 🛠️ Zero-Token Technical Health Audit: Instant assessment of robots.txt AI search bot whitelisting (PerplexityBot, ClaudeBot, GPTBot, OAI-SearchBot, Google-Extended), Schema.org entity integrity, and title hierarchies.
• 📑 Standard /llms.txt & /llms-full.md Generator: Auto-crawl website navigation and produce high-density, markdown-first documentation designed for LLM RAG agents.
• 🏷️ Schema.org & FAQ Synthesis: Automatically extract buyer Q&As and generate compliant JSON-LD structured data snippet patches for immediate CMS injection.
• ⚡ KDD 2024 Citability & Retrieval Quality: Audit page Text-to-HTML ratio, client-side rendering (CSR) risks, canonical compliance, and content freshness signals.
• 🔌 Native MCP Server (8 Tools): Connect seamlessly with Cursor, Claude Code, and Codex to diagnose and fix GEO issues during active development.

---

⚖️ Community Edition vs Commercial SaaS (geofn.com)

open-geo is built on an Open-Core architecture. The community edition provides free, zero-marginal-cost technical diagnostic tools for individual developers. For enterprise teams and digital agencies requiring multi-model concurrent benchmarks and continuous monitoring, visit geofn.com.

| Dimension | 🌐 open-geo (MIT Open Source) | 🔒 geofn.com (Commercial SaaS) | | :--- | :--- | :--- | | Technical Audit | ✅ Single-URL instant static audit & patches | 🔒 Full-site automated crawl & dynamic property drift alerts | | AI Visibility Probing | ✅ Single-model AI visibility probing (Developer-friendly, offline fallback) | 🔒 4-Engine Concurrent Matrix (ChatGPT + Perplexity + Google AIO + Gemini) with managed proxy pool & zero API setup | | Remediation | ✅ Standardized static templates (robots.txt, JSON-LD) | 🔒 Dynamic Competitor Entity Gap Diff ("Why AI cited your competitor instead of you") | | Monitoring | ✅ Point-in-time single run snapshot | 🔒 24/7 AI Citation Drift Tracking with Slack/Email alerts | | Fulfillment | ✅ Developer CLI terminal & raw Markdown | 🔒 White-label Agency Portal & Executive Pitch PDF Proposals |

👉 Learn more about Commercial SaaS on geofn.com →

---

🔌 Using with Cursor & Claude Code (MCP)

Recommended: install globally for instant startup:

npm install -g open-geo

Add to your claude_desktop_config.json or Cursor MCP settings:

{
  "mcpServers": {
    "open-geo": {
      "command": "open-geo",
      "args": ["mcp"]
    }
  }
}

Or via npx without installation:

{
  "mcpServers": {
    "open-geo": {
      "command": "npx",
      "args": ["-y", "open-geo", "mcp"]
    }
  }
}

Once connected, your AI assistant can invoke all 8 native MCP tools:

• geo_diagnose_failure: Diagnose root-cause GEO failures (AI crawler blockades, missing discovery standards, entity fragmentation) and generate actionable code patches.
• geo_audit_webpage: Comprehensive technical readiness audit for HTML structure, Schema.org, llms.txt, and AI crawler whitelisting.
• geo_generate_llms: Crawl site navigation and generate standard /llms.txt and /llms-full.md for LLM RAG agents.
• geo_generate_faq: Extract core Q&As and generate Schema.org FAQPage JSON-LD from URL or raw text.
• geo_retrieval_quality: Audit page Text-to-HTML ratio, client-side rendering (CSR) risk, and Canonical tag compliance.
• geo_content_freshness: Audit page publishing date, modification timestamp, and Schema.org dateModified freshness signals.
• geo_simulate_score: Zero-token local heuristic estimation of overall GEO score (0-100) synthesizing discovery stack, robots accessibility, citability, and entity consistency.
• geo_suggest_keywords: Mine real-time Google & Bing search suggestions and intent stages (ToFU/MoFU/BoFU).

---

❓ FAQ

What is open-geo? An open-source AI visibility checker and GEO (Generative Engine Optimization) toolkit. It audits your robots.txt AI-crawler rules, llms.txt, and Schema.org entities, probes how AI models respond to your brand, and generates patches — running 100% locally.

Is there an open-source alternative to OtterlyAI, Profound, or Peec AI? Yes — that's the gap open-geo fills. Unlike those closed SaaS dashboards, open-geo is MIT-licensed and self-hosted: your data never leaves your machine, and every rule it applies is auditable.

Which AI models can I probe? open-geo probe runs single-model probes with your own API key — OpenRouter (100+ models with one key) or direct Gemini / DeepSeek keys. The hosted dashboard at geofn.com runs multi-engine monitoring (ChatGPT, Perplexity, Google AI Overviews, Gemini).

Does open-geo upload my data? No. Everything runs locally with zero telemetry and zero data uploaded.

How is it different from a visibility dashboard? Dashboards observe; open-geo fixes. It generates executable patches for robots.txt, llms.txt, and JSON-LD, and ships MCP tools so coding agents like Cursor and Claude Code can apply them for you.

🤝 Contributing & License

Contributions are welcome! Please feel free to submit a Pull Request.

This project is licensed under the MIT License - see the LICENSE file for details.