Picking an llms.txt tool depends on your stack, not on which is "best".

Best for: WordPress sites already on AIOSEO

llms.txt generation is built into AIOSEO and switched on by default, including the free Lite version. llms-full.txt and the markdown settings need Pro. Check the output before you rely on it: Common Crawl's study of its July 2026 crawl found AIOSEO files never included the one-line summary (0.00% of 73,136 files) and listed a median of 138 links, so they read more like a sitemap than a curated guide. That is one crawl sample, not a verdict on the plugin.

Best for: WordPress sites already on Yoast, and Shopify stores already on the Yoast app

Free in the regular WordPress plugin, no Premium upgrade needed. Yoast SEO for Shopify added llms.txt in March 2026, but that app is paid: from $19 USD per 30 days ex VAT, with a 14-day trial. Shopify now serves /llms.txt and /llms-full.txt on every store by default (since 28 May 2026), so you no longer need an app just to have the file. In Common Crawl's July 2026 crawl, 82.30% of Yoast files included the one-line summary.

Best for: WordPress sites on Rank Math

Equivalent toggle inside the Rank Math plugin. Slightly more configurable than Yoast on which post types feed the file. If Rank Math is already installed, do not bolt on a second generator.

Best for: A quick free draft built from your site navigation

The free generator pulls links from the part of the page you pick (header, nav, footer or all) and gives you a file in seconds. It does not choose your best pages for you, so curate the result. The paid Business+ plan is $879 a month billed yearly or $1,100 billed monthly (UK visitors see EUR 799 and EUR 999 by default). That is only worth it for enterprise content teams.

Best for: WordPress sites that need both files and are not on AIOSEO / Yoast / Rank Math

One of the few WP plugins that ships llms-full.txt as well as llms.txt. Worth knowing if you want both files and your SEO plugin does not make llms-full.txt for free. How much AI agents use llms-full.txt is unclear (see the fetch data above).

Best for: Hosted documentation, zero-config setup

Auto-generates llms.txt, llms-full.txt and a .md version of every page with no config. If your docs are already on Mintlify, there is literally nothing to do. Mintlify also publish the strongest comparison piece on the wider docs-platform landscape.

Best for: API documentation where you need bot-specific delivery

Generates llms.txt, detects LLM bot traffic and serves markdown instead of HTML to those bots (Fern's own figure is a 90%+ token cut). Content tags like <llms-only> and <llms-ignore> give per-block control, and its analytics break traffic down by LLM provider. Fern stopped generating llms-full.txt on 12 June 2026, saying it saw little use next to llms.txt plus per-page markdown. Pricing is now a free plan and custom Enterprise, with no $400 tier.

Best for: Teams already on GitBook for docs

Publishes both files automatically. No granular content controls, no LLM-specific analytics. Solid low-friction option if GitBook is where the docs already live.

Best for: Interactive API docs with basic llms.txt

On by default for every project and included on every plan. Append .md to any page URL for a markdown version. ReadMe's docs do not mention llms-full.txt, and it offers no AI-specific analytics. Fine if you already use ReadMe; not a reason to switch to it.

Best for: OpenAPI-first docs

Creates llms.txt by default, and you can turn it off. No llms-full.txt. Worth knowing: Redocly have publicly argued llms.txt is overhyped, repeated it in September 2026 after a year's wait, and are putting their effort into MCP server support instead. Picked because of OpenAPI strength, not because of AI features. If AI discoverability is your reason for choosing a docs platform, Mintlify, GitBook or Scalar all do more.

Best for: OpenAPI-first teams publishing on Scalar Docs

Generates interactive API references from OpenAPI specs and has tooling for converting them to LLM-readable markdown. Every published Scalar Docs site now serves /llms.txt and /llms-full.txt automatically. You do not write or configure them.

Best for: Next.js docs where you own the code

Docs framework for Next.js. You can add middleware that detects AI agents and serves markdown. Running npx @fumadocs/cli feature llms now creates the llms.txt, llms-full.txt and per-page markdown routes for you. The code then lives in your project, so you own and maintain it. Full control, much less setup than before.

Best for: Docusaurus-based docs sites

Generates llms.txt, llms-full.txt and per-page .md files. Batch processing handles large sites. Install via npm, configure in docusaurus.config.js. The most complete of the three community plugins.

Best for: Astro Starlight documentation

Generates llms.txt, llms-full.txt and a llms-small.txt for cheaper token contexts. Per-page .md exposure. Active, well-maintained plugin.

Best for: Python (MkDocs) or Vue (VitePress) docs

mkdocs-llmstxt (llms-full.txt through its full_output setting) and vitepress-plugin-llms both generate llms.txt and llms-full.txt during build. Hugo has no dedicated plugin but its output format system makes it a 30-minute custom config job. Eleventy (eleventy-plugin-llms-txt) and Gatsby (gatsby-plugin-llms-txt) have plugins too, but both are llms.txt only.

Best for: Python projects documenting with Sphinx

sphinx-llms-txt generates llms.txt and llms-full.txt during your normal Sphinx build with zero configuration. If the docs are hosted on Read the Docs, the platform will serve those files, but it does not create them, so you still need the plugin or a hand-written file. The lowest-effort option on this whole list for Python docs.

Best for: Laravel and PHP apps wanting programmatic generation

stolt/llms-txt-php. Use when llms.txt needs to update on content changes (CMS-driven sites, headless setups) rather than being a static file. It now generates llms-full.txt as well. Stable, well-tested, low ceremony.

Best for: Quick one-off llms.txt from any URL, if you already use Firecrawl

Still online, but deprecated: Firecrawl stopped maintaining it after 30 June 2025 and points people to its main API. The page asks for a Firecrawl API key for full generation. Through the API it costs 1 credit per URL, crawls up to 100 pages (10 by default) and only builds llms-full.txt if you ask for it. Expect to hand-edit the output (auto-crawls always over-include).

Best for: Quick browser-based draft from a URL

No signup, paste a URL, edit in-browser, download. Output is a starter, not a finished file. Use it the same way you would use Firecrawl - draft, then curate.

Best for: Sitemap-driven generation with a built-in validator

Built by Jan Wilmake with Parallel's support. Generates the file from your sitemap rather than a crawl, and bundles a validator plus an MCP converter in the same toolkit. The sitemap approach means it inherits whatever curation your sitemap already has - often a better starting draft than a raw crawl.

Best for: Catching malformed entries before you deploy

Independent validator at llmstxtvalidator.dev. Flags unsupported directives, broken absolute URLs, missing required fields. One catch: it currently applies llms.txt rules to llms-full.txt too, so read its warnings on that file with care. Always run a generated file through this before shipping it - especially the WordPress plugin outputs, which drift on edge cases.

Best for: Checking the schema and structure of the pages your llms.txt points at

Not an llms.txt generator. Validates the schema and structure of the pages your llms.txt points at. Google's July 2026 guide says optimising for AI features in Search is still SEO and needs no special markup. Useful pair with whichever generator you pick.