GeoScan — does AI cite your website?.

geoscan.seo7.es is a free SEO7 tool: enter your site address and in 10 seconds you learn how well AI (ChatGPT, Gemini, Perplexity, Claude, Copilot) can find, understand and cite your page. It analyzes 305 signals across 7 categories, plus a real query to a language model that shows whether AI names your brand or a competitor. The on-screen result is free, the full report with a PDF arrives by email. A studio lead magnet: leads land in the CRM.

Client: SEO7 own project · Site: https://geoscan.seo7.es · Category: GEO / AI tool · Price for the user: €0 (free)

Tech stack

• Frontend: a single page in plain HTML / CSS / Vanilla JavaScript, no frameworks.
• Orchestration: self-hosted n8n (Docker), two workflows: a synchronous scan and a report with a lead.
• LLM: DeepSeek for niche detection and the Reality axis.
• PDF: a separate Node + pdfmake service (Cyrillic) under systemd: 7 category tables, a 2×2 matrix and a Reality block.
• Infrastructure: nginx (vhost, TLS via certbot, security headers, gzip, cache) on the VPS next to the main seo7 backend (Node + MySQL).
• i18n: three locales ES / RU / EN, formal voice, results written rather than machine-translated.

Full feature list

🌐 Interface and analysis

• Check by a single URL, no registration or install
• Hero canvas visualization: 305 check names in seven clusters with connecting sparks
• Rotating AI names in the headline
• Three locales (ES / RU / EN) with the result in the user's language
• Cookie notice, localized links, responsive layout with a dedicated mobile design

🧠 Scoring engine

• 305 signals across 7 categories, weighted by impact
• GEO Readiness Score from 0 to 100 and an A–F grade
• Core / non-core scoring: what does not apply is not penalized, an honest 100 is reachable
• 2×2 matrix: technical readiness × real citation

🤖 Reality axis

• A real query to a language model about the site's niche
• Checks whether AI names the brand or its competitors
• Returns the actual competing domains

📄 Report

• Score, per-category breakdown and matrix on screen, free
• Full report with every signal and a PDF by email
• PDF: 7 category tables, a 2×2 matrix and a Reality block

🔗 Integrations

• Lead straight into the studio CRM (/admin/crm) with the URL, the score and the competitors
• Telegram and email notifications for every lead
• A methodology repository on GitHub (high-level method, no source, CC BY 4.0)

🛠 Infrastructure and deployment

• Synchronous scan in one pass, no Redis or queues
• URL parsing with regular expressions (the n8n sandbox has no global URL)
• nginx with TLS, security headers, gzip and asset versioning

Process

1. Concept — defined two axes: the page's technical readiness and real citation by AI. 2. Scoring engine — a registry of 305 checks across 7 categories, core / non-core logic for an honest 100. 3. Reality axis — wired in the DeepSeek query: niche, the citation fact and real competitors. 4. Frontend and i18n — a single page in Vanilla JS, the hero visualization, three locales. 5. Report and PDF — a free on-screen result, a full report with a PDF by email. 6. Integrations — leads into the studio CRM, notifications, a methodology repo on GitHub. 7. Deployment — a synchronous n8n workflow, nginx with TLS on the production VPS.

Results