llms.txt and llms-full.txt
llms.txt gives supported AI tools a compact map of your site: an H1, a one-line summary, and grouped links. As of 2026, geodeck generates /llms.txt for every verified domain, plus /llms-full.txt as a clean full-page copy capped at 50,000 bytes. Google's own documentation says llms.txt is not required for Google AI Search, but some other tools can use cleaner page maps when they exist.
What does the llms.txt spec require?
We follow the llmstxt.org proposal: an H1 with the site name, a short summary, then grouped lists of important links. It is optional, but useful for tools that choose to read it.
What does a real llms.txt file look like?
# Acme Corp > Acme is a B2B SaaS that ships logistics dashboards. ## Product - [Pricing](https://acme.com/pricing): Two plans, starts at €19/month. - [API](https://acme.com/api): REST + integrations. ## Resources - [Docs](https://acme.com/docs) - [Changelog](https://acme.com/changelog)
How quickly does geodeck refresh the files?
Files rebuild on every webhook event and on the sitemap diff cron (daily). Responses are cached at the edge for up to 5 minutes (public CDN) or 10 minutes (shared cache).
What is the clean page copy?
Each tracked URL is saved as clean text in /llms-full.txt so supported AI tools can read the page without layout noise.
How do I get a 100/100 llms.txt score?
geodeck generates your llms.txt automatically. Getting a perfect score means making sure the generated file is complete and well-structured. Here is exactly what to do:
- Verify your domain. Without a verified domain, geodeck cannot publish your llms.txt. Go to your dashboard, add a domain, and follow the DNS TXT verification steps. The file appears at geodeck.icu/api/llms/yourdomain.com once verified.
- Import your sitemap or add URLs. Go to the domain detail page → Index tab → Import Sitemap. Paste your sitemap URL and import. geodeck needs at least 15 URL entries to fill the llms.txt sections and earn the full list-item score.
- Add page titles and summaries. Each URL in the llms.txt shows its title and a one-line summary. geodeck pulls these from your page metadata. Ensure every tracked page has a unique <title> tag and a meta description (50-160 characters). Missing titles appear as blank entries and reduce file quality.
- Enable the Markdown Mirror (Agency plan). The llms-full.txt file earns 10 extra points. It is populated when geodeck mirrors your page content as Markdown. This happens automatically via the webhook integration or when you trigger a sitemap import. On the Agency plan, the full mirror is enabled by default.
- Keep the file size in range. The generated file should be between 200 bytes and 50,000 bytes to earn the size bonus. A file with 15-200 URLs and reasonable summaries hits this range naturally. If your sitemap has thousands of URLs, geodeck samples the most recent ones to stay in range.
Score 100/100: domain verified (llms.txt accessible), H1 present, blockquote summary present, 3+ sections, 15+ list items, llms-full.txt accessible, file size 200-50,000 bytes.
How is the llms.txt score calculated?
The score is 0-100 and measures how complete and well-structured your llms.txt file is. Each signal adds points:
- File found and accessible (+20 pts) - the most important signal. Without it, the score is 0.
- H1 heading with your site name (+15 pts) - geodeck generates this from your domain name automatically.
- Blockquote summary paragraph (+15 pts) - geodeck uses your homepage meta description. Ensure it is present and 50-160 characters.
- At least 1 section (+10 pts) - a section is an H2 heading grouping related links.
- 3 or more sections (+10 pts) - geodeck groups your URLs by category when you have enough pages.
- 5 or more list items (+10 pts) - each URL is a list item. Import at least 5 pages.
- 15 or more list items (+5 pts) - import 15 or more pages for full coverage.
- llms-full.txt accessible (+10 pts) - the clean full-page copy. Available on Agency plan or when webhook integration is active.
Does llms.txt actually matter for AI search?
Google's AI features documentation is explicit that no special file is required, and schema.org's getting-started guide takes the opposite approach of asking you to mark up the page itself rather than publish a side file. The GEO paper that coined generative engine optimization used a benchmark of real queries to test what actually moves visibility, rather than assuming any one file format helps. llms.txt is plain Markdown, so it costs almost nothing to keep alongside whichever approach wins.
You don't need to create new machine readable files, AI text files, or markup to appear in these features.
— Google Search Central documentation, “AI features and your website”
Schema.org provides a collection of shared vocabularies webmasters can use to mark up their pages in ways that can be understood by the major search engines: Google, Microsoft, Yandex and Yahoo!
— schema.org, “Getting started with schema.org using Microdata”
We facilitate systematic evaluation by introducing GEO-bench, a large-scale benchmark of diverse user queries across multiple domains, along with relevant web sources to answer these queries.
— Aggarwal et al., “GEO: Generative Engine Optimization” (arXiv:2311.09735)
What else can I read?
Keep reading the rest of the geodeck documentation, or jump to the platform pages.
- All docs →Three-step quick start and the full guide list.
- Domain verification →Add a single DNS TXT record so geodeck can manage your site.
- Publish webhook →POST publish, update, and unpublish events with HMAC-SHA256.
- Indexing & ping providers →IndexNow, Google Search Console, Google Indexing API, and XML-RPC pings.
- Visibility score →How the 0-100 score breaks down across crawler access, page details, answer clarity, freshness, and site basics.
- Blog →Guides on Google visibility, AI answers, indexing, and getting found by more customers.
- Scan for free →Get your website found. On Google and AI answers.
- llms.txt Generator →Generate, host, and auto-update your llms.txt file. Give supported AI tools a clean map of your most important pages. Works with any CMS and is included on every geodeck plan.
- AI Visibility →See how ready your site is for AI answers. geodeck checks crawl access, llms.txt, schema, answer clarity, Perplexity citations, Bing AI visibility signals, and Search Console coverage.
- Answer Engine Optimization →Answer Engine Optimization (AEO) is the practice of structuring content so AI systems like ChatGPT and Perplexity choose your pages as the definitive answer. Learn the tactics that work.
- GEO Tools →The best generative engine optimization (GEO) tools for AI visibility readiness, llms.txt, schema, sitemap automation, Perplexity citation tracking, Bing AI visibility signals, and Google Search Console coverage.
- AEO vs GEO vs SEO →AEO, GEO, and SEO all aim to drive traffic - but they optimize for completely different signals. Here's the definitive comparison of Answer Engine Optimization, Generative Engine Optimization, and traditional SEO.