GEO playbook
llms.txt for SEO
Published 11 Aug 2026 · By Duy Nguyen
Search is splitting into two surfaces: classic results and AI answers. Generative Engine Optimization (GEO) is the discipline of being cited by ChatGPT, Perplexity, Gemini and Claude — and llms.txt is one of the cheapest GEO assets available. It is not a ranking factor (nothing AI-side is confirmed), but it is an orientation file that costs minutes to publish and gives AI agents a precise map of your site.
Why does llms.txt matter for AI search?
AI engines cite what they can parse and trust. A navigator file tells them your site's structure in one fetch; citation studies (Victorino LLC, The Digital Bloom, Position Digital, Omniscient Digital) show that structured, entity-rich, question-oriented documents are cited more often. llms.txt operationalizes that research in a single text file — which is exactly what the five GEO checks measure.
What are the five GEO checks?
The generator scores your file against five thresholds drawn from published research: ski-ramp entity placement (≥45% of entities in the first 30% — cited content concentrates there), chunk self-containment (≥80% of sections at 50–150 words — a 2.3× citation lift), question-form headings (≥20% — a 2.8× lift), entity density (≥15% — cited answers average 20.6%), and definitional phrasing (≥1.0 per 100 words — cited at 36.2% vs 20.2%). Each failing check has a one-click fix in the generator.
How do you start with llms.txt in 2026?
Three steps. First, publish the file: generate it from your sitemap with the free generator, review the sections, download, and upload to your domain root. Second, verify access: allow GPTBot, ClaudeBot, PerplexityBot, Google-Extended and OAI-SearchBot in robots.txt (the generator's robots check audits nine crawlers), and keep Cloudflare's "Block AI Crawlers" toggle off. Third, maintain: regenerate whenever your structure changes, and re-score the file after content updates.
What are the limits of llms.txt?
Be honest about adoption: Google ignores it, and a 2026 Ahrefs study found most published llms.txt files received zero requests in a month. llms.txt is early-stage insurance, not a traffic pump. Pair it with the fundamentals — technical SEO, real content, meta tags and JSON-LD (see llms.txt vs meta tags) — and treat the AI layer as a compounding option rather than an overnight win.
Where does llms.txt fit in a GEO strategy?
In a mature GEO setup, llms.txt is the orientation layer: it sits above your sitemap (see llms.txt vs sitemap), complements robots.txt (comparison here), and is backed by llms-full.txt for agents that want the whole corpus in one request (what is llms-full.txt). Publish all four files, keep them honest, and let the AI layer compound while classic SEO does the heavy lifting.