How do I generate an llms.txt file with llmstxt.new?

To generate an llms.txt file on llmstxt.new, open the generator, enter the URL of the website you want to process, and choose either the standard llms.txt or llms-full.txt output. A Firecrawl API key is optional but recommended: without one you can still request a file, while entering a key enables full generation. The page also notes that the underlying API can be called directly, and that this API is being deprecated in favor of the main endpoints after June 30, 2025, with an example repo at github.com/mendableai/create-llmstxt-py.

What the tool does

llmstxt.new is a web-based generator labeled "LLMs.txt generator v2" that produces consolidated text files from websites for LLM training and inference, powered by Firecrawl. In practical terms, you give it a site URL and it returns a text file that describes the site's content in a form an LLM can consume, rather than you assembling that file by hand.

Two output types are offered on the page:

The page does not publish a detailed spec of what each file contains, so treat the distinction above as the working choice the interface presents rather than a guaranteed content contract.

Step-by-step generation

1. Open the generator. Go to llmstxt.new and locate the "LLMs.txt generator v2" section.
2. Enter the target URL. Provide the address of the website you want to convert. The page explicitly asks for a URL before it will generate a file.
3. Decide on an API key. The page says: "Enter Firecrawl API key for full generation" and separately notes that "for a better experience, use an API key from Firecrawl." If you have a Firecrawl key, enter it; if not, you can still attempt generation, but expect the fuller output to depend on the key.
4. Choose the output. Select llms-full.txt if you want the fuller file, or use the standard llms.txt generation path for the lighter file.
5. Generate and check the result. Run the generation and open the returned text. Confirm it covers the pages you expected; if it looks thin, that is the signal to retry with a Firecrawl API key for full generation.

Expected result: a downloadable or copyable text file describing the site, ready to place at your site root or feed to an LLM workflow.

Using the API instead of the page

The page states the generator "can also be called via API," so scripted or repeat generation does not require the web form. The page does not document endpoint paths, parameters, or authentication details beyond the Firecrawl key mention, so consult the linked example repository (github.com/mendableai/create-llmstxt-py) for a working reference implementation before wiring it into a pipeline.

Common snags

• No URL entered. The generator will not produce a file; the URL field is a hard requirement.
• Thin output without a key. The page ties "full generation" to a Firecrawl API key, so a keyless run may give you less than you need.
• Assuming long-term API stability. The page warns the API is being deprecated in favor of main endpoints and will no longer be maintained after June 30, 2025, though the endpoint stays active. If you are building something durable, plan around the main endpoints or the example repo rather than this specific endpoint.
• Expecting a published format spec. The page describes purpose, not a schema; verify the generated file against your own LLM ingestion needs.

Is it the right tool for you?

Use llmstxt.new when you want a quick, no-setup way to produce an llms.txt or llms-full.txt from a site URL and are fine with Firecrawl powering the extraction. Prefer the API or the example repository when you need repeatable, automated generation, and prefer the main Firecrawl endpoints if you need something maintained past June 30, 2025. The page does not state pricing or login requirements for the generator itself, so confirm those on the site before relying on it at scale.