Troubleshooting & Best Practices for LLM Index

1. Endpoint Not Available (404) #

If /wp-json/davix/v1/llm-index or /llm.txt returns a 404:

Checks:

1. Is the module enabled?
   • Go to Davix AI Engine → LLM Index and confirm Enable LLM Index is ON.
2. Permalinks / REST working?
   • Ensure your WordPress REST API is generally working (since LLM JSON uses a REST route as described in the docs).
3. Caching or security plugin interfering?
   • Some security plugins may block certain REST paths; check their logs.

2. Endpoint Returns Empty or Very Few Items #

If the endpoints load but show very little or no content:

Checks:

1. Post Types & Taxonomies:
   • Verify you have selected the correct post types and taxonomies in LLM Index settings.
2. Max Items Too Low:
   • Check Max Items; if it is set very low, only a small number of entries are returned.
3. Per-Post Overrides:
   • If you used Force Exclude heavily in the Davix LLM Index meta boxes, some posts may be deliberately missing.

3. JSON Endpoint Access Issues #

If the JSON endpoint returns a permission error instead of data:

Checks:

1. Public vs Admin-Only:
   • In LLM Index settings, see whether JSON is configured as public or restricted.
2. Logged-In State:
   • If JSON is admin-only, ensure you’re logged into WordPress with a role that has necessary capabilities (like admin).

4. Using the LLM Index with External Tools #

Your plugin description explicitly says:

“LLM Index (JSON endpoint + llm.txt feed) to expose your content to AI tools.”

Best practices based on this:

1. Decide if endpoints should be public:
   • If you plan to integrate external AI tools, set JSON visibility to public or configure your tools to authenticate.
2. Limit to important content:
   • Use post types/taxonomies and the per-post meta box to keep the index focused on content you actually want AI tools to see.
3. Keep summaries concise:
   • Choose a Summary Length that gives enough context but doesn’t overwhelm the consumer with huge blocks of text.

5. Debugging Using Davix Debug Logs #

While the docs primarily talk about Debug Logs for chatbot, forms, auto-reply, and AI Writer, they also mention that the debug system centralizes logs for Davix modules.

If you encounter odd behavior:

1. Go to Davix AI Engine → Debug Logs.
2. Enable relevant logging (if there is a channel that logs indexing operations).
3. Save and reload endpoints.
4. Check the log file for any errors or warnings related to LLM index behavior.

image IMAGE PLACEHOLDER: Screenshot of Debug Logs page where you can enable logging channels and click “View log”. IMAGE image