Category

Markdown and LLM Supported Knowledge Base in BoldDesk

Updated:

As Large Language Models (LLMs) like ChatGPT, Claude, and Copilot become essential for information access and knowledge support, making your documentation LLM-compatible ensures accurate, context-aware responses.

BoldDesk now provides LLM-friendly knowledge base (KB) documentation with Markdown (.md) and llms.txt support for efficient AI-driven interactions.

Why optimize documentation for LLMs?

LLMs rely on structured, machine-readable content to deliver precise answers. Poorly formatted documentation can lead to incomplete or inaccurate responses.

By using Markdown and llms.txt, you enable:

  • Better parsing: AI models read Markdown faster than HTML.
  • Improved context: Structured headings and lists enhance meaning.
  • Efficient retrieval: llms.txt serves as a centralized index for all Markdown pages.

Why use Markdown (.md)?

Markdown is the preferred format for LLMs because it is:

  • Lightweight: Less overhead compared to HTML.
  • Structured: Easy to organize with headings, lists, and emphasis.
  • Parse-friendly: Simpler for AI models to process.

Example use case
Instead of sending an HTML page to the LLM, provide the Markdown version for faster and more accurate interpretation.

Markdown in BoldDesk

BoldDesk automatically makes all KB pages available in Markdown format.

You can:

  • View any KB page in Markdown by adding .md to its URL.
  • Copy Markdown content directly from the Article Details page.

Steps to access Markdown:

  1. Navigate to the Article Details page.

  2. Click Copy page to duplicate the Markdown content.

    View as Markdown.png

  3. Select View as Markdown to open the article in a new tab as Markdown.

    Markdown KB.png

Here’s a sample URL for viewing the page as a Markdown (.md) file: https://yourdomain.com/kb/article/12345/getting_started.md

What is llms.txt?

llms.txt is a proposed standard for exposing web content in text formats optimized for LLMs.

It serves as an index file listing all available Markdown pages on your KB site.

Benefits of llms.txt

  • Centralized discovery: LLMs can quickly find all Markdown pages.
  • Improved AI efficiency: Reduces crawling complexity for AI systems.

Access example
Append /llms.txt to your KB root URL: https://yourdomain.com/llms.txt

llms.txt in BoldDesk

BoldDesk automatically generates llms.txt for KB sites using a custom domain.

Once your domain is mapped:

  • The file is available at /llms.txt.
  • It includes all published KB articles in Markdown format.

Example: Visit BoldDesk Support file to access BoldDesk documentation site.

What is llms-full.txt?

llms-full.txt file provides a compact directory of text-friendly pages intended for use by Large Language Models.

It lists the main Markdown resources available on the site, making it easy for LLMs to locate and read structured content.

Benefits of llms-full.txt

  • Faster content discovery: LLMs can quickly map your site’s structure without scanning every page.
  • Consistency: Helps maintain a predictable, organized structure for AI-friendly content.

Access example
Append /llms-full.txt to your KB root URL. It should appear as follows: https://yourdomain.com/llms-full.txt

llms-full.txt in BoldDesk

BoldDesk automatically generates llms-full.txt for KB sites using a custom domain.

Once your domain is mapped:

  • The file is available at /llms-full.txt.
  • It includes all published KB articles in Markdown format.

Example: Visit BoldDesk Support file to access BoldDesk documentation site.

Frequently Asked Questions (FAQs)

  1. Why is Markdown better than HTML for LLMs?
    Markdown is lightweight and structured, making it easier for AI models to parse and interpret compared to complex HTML.
  2. Do I need to create llms.txt manually?
    No. BoldDesk automatically generates llms.txt for KB sites with custom domains.
  3. Can I use this feature without a custom domain?
    llms.txt is only available for KB sites mapped to a custom domain.
Was this article useful?
Like
Dislike
Help us improve this page
Please provide feedback or comments
Comments (0)
Access denied
Access denied
Access denied
Access denied

No articles or sections found
No articles or sections found