1 min read
|
Saved February 14, 2026
|
Copied!
Do you care about this?
The article discusses the need to optimize websites for AI agents, which prefer structured content like Markdown. It explains how the author modified their blog to serve Markdown files alongside HTML, using a GitHub Actions workflow for deployment. The piece raises questions about content usage by AI but suggests it may lead to better attribution.
If you do, here's more
Dries Buytaert highlights the emergence of AI agents as a new audience for websites, alongside humans and search engines. Most sites are not yet optimized for these AI agents, which prefer clean, structured content. Markdown is presented as an ideal format because it is readable and semantic, lacking unnecessary navigation elements. Buytaert implemented a solution on his blog that allows AI agents to access the source Markdown files directly alongside the HTML versions.
To make this work, he set up a process using GitHub Actions since his GitHub Pages blog does not support custom Jekyll plugins. This workflow builds the Jekyll site and ensures that Markdown files are copied to the right locations before deployment. For each blog post, it extracts the date and slug from the filename and creates a corresponding Markdown file accessible at a specific URL.
Buytaert raises a concern about whether this approach aids AI companies at the expense of website traffic. However, he argues that providing cleaner content could improve attribution and lead to more accurate AI responses that reference original works. He sees this as a natural progression for the web, emphasizing the importance of making information accessible.
Questions about this article
No questions yet.