llms.txt Generator

How to use

1. 1Pick a preset that matches your site (SaaS, open source, docs site, personal site) or start from scratch.
2. 2Fill in the title and one-line summary at the top. The summary becomes a > blockquote so LLMs can grab it in one read.
3. 3Add sections with the +Add another section button. Name a section Optional to mark lower-priority links per the spec.
4. 4Inside each section, add link rows with text, an https URL, and an optional one-line description.
5. 5Watch the live preview and the Validation panel. Fix any flagged errors, then click Copy or Download to save the file as /llms.txt at the root of your domain.

About this tool

llms.txt Generator builds a valid /llms.txt file from structured input. Set the H1 title, a one-sentence summary that renders as a > blockquote, optional context paragraphs, and any number of ## sections each containing curated link bullets in the [text](url): description form. A live byte and line count sit next to the preview, a Copy button copies the raw markdown, and a one-click Download saves it as llms.txt with the right MIME type. Browser-side validation flags the mistakes that quietly break adoption: a missing H1, multi-line title, missing URL on a link row, a URL that does not parse as http or https, a duplicate URL across sections, a section with no name, and an empty summary that LLMs will struggle to pick out at a glance. Four ready-made presets cover the workflows site owners ask for most often: SaaS product (marketing, docs, API, policies), Open source project (readme, API docs, examples, contributing), Docs site (guides, reference), and Personal site (about, posts, projects, contact). Each preset seeds the right shape so you only need to swap in your own URLs. Sections can be reordered with Up and Down controls, and naming a section Optional is the spec's convention for lower-priority links that an LLM should fetch only if it has budget. Useful for anyone publishing a docs portal, marketing site, blog, API reference, or open-source library who wants to control what ChatGPT, Claude, Perplexity, Gemini, and the next wave of LLM clients read when they look up your project. Parsing, validation, and output assembly all run in your browser, so the URLs and descriptions you draft stay on your device.

Free to use. Works in your browser. No signup, no login.