llms.txt builder for AI agents
Generate an llms.txt and llms-full.txt file for your site, following the standard proposed by Jeremy Howard in 2024 and adopted in 2026 by major LLM agents (ChatGPT, Claude, Perplexity, Gemini). The file is the AI-friendly counterpart to sitemap.xml and robots.txt: it tells AI agents which pages on your site matter, in what order, and how to read them, so they can answer user questions with high-quality, current information from your domain instead of guessing or scraping noise.
Docs / primary pages
Examples / use cases
Optional / nice-to-have
Upload the generated file to https://yourdomain.com/llms.txt. The longer llms-full.txt is optional; use it when you want to bundle full documentation content rather than just links.
What is llms.txt and why it matters in 2026
llms.txt is a plain Markdown file proposed by Jeremy Howard in late 2024 and quickly adopted across the AI ecosystem in 2025-2026. It sits at the root of a domain, exactly like robots.txt, but its purpose is the opposite: instead of telling crawlers what to avoid, it tells AI agents which pages on the site matter for answering questions about the project. A typical file lists a project name, a one-line summary, a longer context paragraph, then a small set of links grouped under headings like Docs, Examples and Optional. When an AI assistant is asked about your product, it can read this short Markdown file in one fetch and know exactly where to go for the canonical answer rather than scraping a noisy homepage or out-of-date Common Crawl snapshot.
The format matters because language models read tokens, not HTML. A 30-line Markdown file with ten curated links is far cheaper to ingest than crawling 200 pages and trying to figure out which is authoritative. Major AI clients including ChatGPT browsing mode, Perplexity, Claude with web access, and several open-source agent frameworks now check for /llms.txt as part of their answer pipeline. Sites that publish a clean file get cited more often and with more accurate context than sites that do not.
How this llms.txt generator works
The page builds the file in your browser as you fill the form. The required pieces are the project name, the site URL, a one-line summary, and a free-form context paragraph. The optional pieces are three grouped link lists: Docs (primary documentation pages), Examples (concrete tutorials or use cases), and Optional (anything that helps but is not central). Each link has a short title that becomes the Markdown anchor and a one-line description that the AI agent will read alongside the URL. As you edit, the generator re-renders the file in real time and the Best-practice checks tab flags missing pieces, links that are too long, descriptions that look duplicated, and the rough token cost the file will have when an agent ingests it.
- Start with the H1 line: a single hash followed by the project name. The file MUST start with this line.
- Add the summary: a blockquote (line starting with
>) holds the one-line description used by agents to decide if the file is relevant. - Add context: one or two plain Markdown paragraphs that explain what the project does and what kind of questions the file answers.
- Group links under H2 headings: Docs, Examples, Optional, or your own categories. Each line is a Markdown bullet with a Markdown link and a colon-separated description.
- Optionally publish llms-full.txt: a longer file with the actual documentation content inlined, useful when you want the agent to answer without further fetches.
Common use cases for publishing llms.txt
- SaaS documentation site. Surface your “Getting started”, “Reference”, “Pricing” and “Support” entry points so AI assistants can answer “how do I do X with your product” without crawling marketing pages.
- API publisher. Point AI agents to the OpenAPI spec, the SDK docs and the changelog. Agents that help developers integrate the API will find the right entry point in one fetch.
- Open-source library. List the README, the CONTRIBUTING guide, the API reference and the FAQ. AI coding assistants like Claude Code and ChatGPT Code Interpreter use this to answer “how do I install” and “what version is current”.
- Knowledge base or help centre. Surface the categories AI agents should prioritise (refunds, account, security) so customer-support chatbots ground their answers on your authoritative content instead of guessing.
- Tools hub like PeopleAreGeek. Group tools by hub (Network, SEO, Cyber) so an AI agent answering “is my DNS propagated” can point the user directly to the propagation checker.
- Personal portfolio or blog. List your most cited articles and your bio. AI agents asked “who is <your name>” will use this instead of scraping a Twitter bio.
Limitations and adoption notes
llms.txt is a young convention. Adoption is growing fast in 2026 but not all AI agents check for the file yet, and there is no enforcement mechanism: an agent can ignore it just as it can ignore robots.txt. The file is best treated as a high-value invitation, not a hard contract. The format is intentionally tiny (most files fit under 100 lines) so the cost of publishing it is also tiny: a single Markdown file uploaded to the root of the domain. Keep it updated when your documentation structure changes, otherwise agents will point users to broken or stale anchors. The longer llms-full.txt is optional and trades file size for fewer follow-up fetches; only publish it if you actually want agents to consume long content in one shot.
The tool runs in your browser. Your project name, summary, paragraph and link list are processed locally; nothing is sent to PeopleAreGeek or to a third party. The Best-practice checks tab estimates the token cost of the file as it would be ingested by a large language model, so you can keep the file lean.
Frequently asked questions
How is llms.txt different from robots.txt and sitemap.xml?
robots.txt tells crawlers which paths to avoid. sitemap.xml lists every URL on the site for search engines. llms.txt is a curated, human-written Markdown summary of the most important pages for AI agents. It is shorter, opinionated and meant to be read, not just parsed.
Do I need llms.txt if I already have a sitemap?
A sitemap helps Google index your site for search. llms.txt helps AI agents answer questions about your project without crawling everything. They serve different purposes; most teams publish both.
Where do I upload the generated file?
Upload it to the root of your domain at https://yourdomain.com/llms.txt. Test by fetching that URL with curl. Some agents also check https://yourdomain.com/llms-full.txt for the longer version.
Should the file include every page on my site?
No. llms.txt is curated, not exhaustive. Include the entry points an AI agent needs to answer the most common questions about your project; for everything else, sitemap.xml is the right place.
What is in llms-full.txt that is not in llms.txt?
llms-full.txt is the optional longer variant that inlines the actual content of your documentation pages, so an AI agent can answer detailed questions without follow-up fetches. It is useful when answers depend on long content and you want agents to consume it in one read.
Will AI agents always honour what I put in llms.txt?
No more than they always honour robots.txt. Major agents (Perplexity, ChatGPT, Claude, Gemini) read it; smaller or anonymous scrapers may not. Treat it as a strong invitation, not as enforcement.
