/llms.txt Generator
Pricing
Pay per usage
/llms.txt Generator
The /llms.txt Generator 🕸️📄 extracts website content to create an llms.txt file for AI apps 🤖✨ like LLM fine-tuning and indexing. Output is available 📥 in the Key-Value Store for easy download and integration into workflows. 🚀
Pricing
Pay per usage
Rating
5.0
(4)
Developer
Jakub Kopecký
Maintained by CommunityActor stats
7
Bookmarked
810
Total users
0
Monthly active users
a day ago
Last modified
Categories
Share
🗂️ /llms.txt Generator
Turn any documentation site into an llms.txt file in a few seconds. Paste a URL and it crawls the site, curates the pages, and writes a Markdown index — plus an optional llms-full.txt with the full content of every page — that AI coding agents can read instead of your whole site.
🌟 What is llms.txt?
A Markdown file that tells AI tools where a website's useful content is: a title, a one-line summary, and sections of links with short descriptions. Coding agents and AI IDEs (Claude Code, Cursor, Windsurf, GitHub Copilot) read it when pointed at a documentation site. There's no evidence search engines or chat assistants use it for ranking or citations — treat it as context for agents, not SEO.
✨ What it does
- Crawls from your start URL, following redirects, ordered the way the site's own navigation lists pages
- Cleans up titles and descriptions, and drops duplicate pages
- Groups pages into sections, with a spec-compliant
Optionalsection for changelogs, blog posts and old doc versions - Links to a page's Markdown version when the site publishes one, so agents read it faster and cheaper
- Can also write
llms-full.txtwith the full text of every page - Can hand off to an AI model to organize sections by topic instead of URL structure
- Tells you if the site already publishes its own
llms.txt
🚀 How to use it
- Enter the Website address.
- Click Start. Download
llms.txt(andllms-full.txt) from the Output tab when it finishes.
Everything else lives under Advanced settings, with defaults tuned for most documentation sites: up to 100 pages, llms-full.txt on, AI organizing off.
Output example
This is the llms.txt generated for https://docs.apify.com/cli/docs, shortened:
# Apify CLI overview | CLI | Apify Documentation> An introduction to Apify CLI, a command-line interface for creating, developing, building, and running Apify Actors and managing the Apify cloud platform.## Docs- [Quick start](https://docs.apify.com/cli/docs/quick-start.md): Learn how to create, run, and deploy Actors using Apify CLI.- [Installation](https://docs.apify.com/cli/docs/installation.md): Learn how to install Apify CLI using installation scripts, Homebrew, or NPM.## Optional- [Changelog](https://docs.apify.com/cli/docs/changelog.md): All notable changes to this project will be documented in this file.
The run also saves a dataset item with links to both files, the crawl stats, and the site's existing llms.txt URL if it has one.
🤖 AI curation (optional)
Turn on Organize with AI to have an LLM group pages by what a developer wants to do ("Getting started", "API reference") instead of URL structure, and write a real summary. It scales to large sites by batching the request, costs about $0.001 per 100 pages on a paid Apify plan (10x on the free plan), and never pushes the run past its timeout — if it fails or runs out of time, you still get the regular URL-based file.
⚠️ Limitations
- No JavaScript rendering. Sites fully rendered in the browser (e.g. Docsify) fail with a clear explanation instead of an empty file.
- Large sites take longer, but always finish. Bigger crawls (up to 5,000 pages) automatically get more memory; if a run is about to hit its timeout, it saves whatever it has and says so.
- Descriptions come from the site. Pages without a meta description get a title only.
- Treat the output as a first draft. Review it, trim what your users don't need, before publishing at
https://your-site.com/llms.txt.
💡 Tips
- Start from the docs root, not the marketing homepage — documentation pages have better titles and descriptions.
- Use Skip pages to leave out old versions, translations or generated API pages.
- Only enable the proxy if the site blocks the crawler — most documentation sites don't need it, and skipping it is faster.