AI Context file

[Daxcor69]

Hello,

I found using different ai services to help me with caddy config were not very helpful. It would be extremely helpful if there was a specific url/file that can be pointed to in an ai request to provide optimized and relevant context to help configure caddy. Here is what a chat gpt session would look like.

I am configuring caddy v2 and I need some help with the configuration. You can find precise context for your responses at https://caddy.com/ai/v2/context.txt.  Please ensure your responses rely heavily on this context as it has been designed specifically to help ai answer caddy questions.

The text file would have a natural language presentation of caddy. Things like an introduction, intended audiences etc followed by the documentation. The documentation doesn't need to be human friendly. You don't have to spend time to help people understand the concepts. This would be more about just the concept. The ai model would make it human friendly for you.

You can also talk about different common use cases and the configuration pattern that makes the most sense.

You can have a section on specifically how to configure for web sockets and long running connections.

You can have a paragraphs on common pinch points and how to handle them.

I hope you get the idea.

Brad foundryserver.com

[notramo]

This specification might be relevant: https://llmstxt.org/

[mholt]

Thanks for the issue! We'll consider this for our next round of website improvements.