geo-site-audit: An Automated SEO & GEO Audit Tool to Track Brand Visibility in AI Search Engines

geo-site-audit is an open-source automated website audit tool designed to address new challenges in brand digital visibility in the AI era. It integrates functions such as traditional SEO basic indicator checks, Core Web Vitals performance testing, brand visibility tracking across 6 major LLMs, and can generate real-time reports deployable to Vercel, helping enterprises understand their position in the AI search ecosystem.

As AI assistants like ChatGPT, Claude, and Gemini become primary channels for information acquisition, traditional SEO can no longer fully cover brand visibility needs. GEO (Generative Engine Optimization) has emerged as a new battlefield, and the geo-site-audit tool was developed to not only check traditional SEO metrics but also automatically detect brand recommendations across multiple LLMs, addressing enterprises' pain points in visibility within the AI ecosystem.

The tool covers six detection dimensions: 1. Traditional SEO audit (basic elements like titles and meta descriptions); 2. Core Web Vitals performance testing (obtaining LCP, CLS, INP data via Google PageSpeed Insights); 3. GEO brand visibility detection (sending 7 brand-related queries to 6 LLMs, calculating GEO score based on AI mention rate + content quality score); 4. Competitor monitoring (scanning AI responses for mentions of competitors); 5. Broken link detection (identifying internal/external dead links); 6. Content gap analysis (marking missing important pages such as FAQs).

The audit process is completed collaboratively by Python scripts: 1. audit.py (Firecrawl crawls pages to extract data); 2. check_ai_visibility.py (OpenRouter calls LLMs to record brand mentions); 3. analyse.py (analyzes SEO issues, performance, etc.); 4. generate_ai_insights.py (Claude Sonnet generates optimization suggestions and roadmaps); 5. generate_report.py (compiles data for front-end display). The front-end is based on Next.js and can be deployed to Vercel for real-time reports.

Deployment requires preparing Firecrawl API Key (for crawling), OpenRouter API Key (for LLM calls), and PageSpeed Insights API Key (for performance data). Configuration steps: Modify _system/config.json to set target URLs, _system/pages.json to define audit pages, _system/check_ai_visibility.py to update brand and competitor lists, then run the audit script.

Marketing teams can gain AI visibility insights; technical teams can monitor website health (Core Web Vitals, broken links); competitor analysis features help track dominant opponents in AI recommendations. The tool fills the gap where traditional SEO tools cannot cover visibility in AI assistants.

The tool is a product of internal projects; the public version is not updated frequently and requires technical capabilities for customization. The cost per LLM API call is low, but large-scale/high-frequency audits need to consider costs.

geo-site-audit represents a new idea from traditional SEO to AI engine optimization. In today's era where AI assistants are information entry points, brand AI visibility is as important as traditional SEO. The tool raises a key question: When users ask AI about the best brands, is your brand mentioned? It provides a reference for enterprises to maintain competitiveness in the AI era.