Our team recently released support for /llms.txt and /llms-full.txt, meaning that Mintlify now automatically generates and hosts all docs in a plain text format for easier ingestion by large language models (LLMs).
In September 2024, Jeremy Howard, Co-Founder of Answer.AI, proposed the /llms.txt file format as a standardized way for websites to provide key information to LLMs. This markdown-based file distills a site's content into a concise, structured format, bypassing the need for LLMs to parse complex HTML, ads, or JavaScript.
By adopting this standard, we're taking a step toward a future where every company provides two versions of their documentation: one for humans and another for LLMs.
Before AI, finding answers from documentation could be a guessing game.
Search within docs hinged on hitting the right keywords. If you looked for “invoice editing” but the docs referred to it as “billing adjustments,” you'd be out of luck. Without effective search, customers often have to resort to deciphering navigation hierarchies and skimming endless pages.
People might try turning to third party search engines to find the right docs or relevant community forums. But this relies on strong SEO and also hitting the right keywords—yet another roll of the dice.
At Mintlify, we've focused on solving these challenges with:
Now, we're providing a better way for LLMs to process documentation, helping people use AI tools to find and apply information about your product:
/llms.txt: AI-friendly navigation at scale
This auto-generated markdown file offers a simplified view of your entire docs navigation, allowing LLMs like ChatGPT or Perplexity to index your content more easily. Think of it as SEO for AI—users can now find product-specific information directly through general-purpose LLMs.
/llms-full.txt: All your docs content in one file
This auto-generated file compiles all your documentation text into a single markdown file.
Some AI tools allow you to paste a link to load information directly into their context window. For example, you can feed your docs into AI coding assistants like Codeium, allowing them to provide context-aware suggestions based on specific details of your product, such as your authentication flow.
Instead of pasting dozens of links to enrich the assistant, you can provide one comprehensive file with /llms-full.txt.
By adopting the /llms.txt standard, we're excited to help lay the foundation for its widespread use across the industry.
These releases are only the start. We're continuing to invest in new ways AI can make documentation more useful and accessible—stay tuned for big updates soon.
If you're interested in future-proofing your docs like Anthropic, Windsurf, Bolt.new, get in touch for a free preview.
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。