add info on llms.txt improvements

Author: devalog

fern/products/docs/pages/seo/llms-txt.mdx

@@ -19,24 +19,21 @@ See an example: [docs.cohere.com/llms.txt](https://docs.cohere.com/llms.txt)
 
 ### llms-full.txt
 - **Comprehensive**: Includes the full content of your documentation.
-- **API-Ready**: Automatically incorporates your full API reference with OpenAPI specifications and SDK code examples (if generated by Fern).
-- **Token efficient**: Removes unnecessary formatting to be as token-efficient as possible.
+- **API-ready**: Automatically incorporates your complete API reference with resolved OpenAPI specifications and SDK code examples for all enabled languages. Each API endpoint page includes its full OpenAPI spec.
+- **Token efficient**: When Fern detects an LLM bot, it automatically serves markdown content from the llms.txt file, reducing token consumption by 90%+.
 
 See an example: [docs.cohere.com/llms-full.txt](https://docs.cohere.com/llms-full.txt)
 
-## Analytics and Monitoring
+## Analytics and monitoring
 
-Fern tracks comprehensive analytics for llms.txt usage to help you understand how AI tools are consuming your documentation. The system monitors performance metrics including load times, processing times, and content length. User agent detection automatically identifies bot traffic versus browser requests, and detailed logging captures request headers for deeper insights.
+Fern tracks comprehensive analytics for llms.txt usage in the Fern Dashboard, including:
+- LLM traffic by provider (Claude, ChatGPT, Cursor, etc.)
+- Traffic breakdown for .md and llms.txt files by agents and humans
 
-These analytics are available in the Fern dashboard, where you can view dedicated tables for llms.txt and markdown file access, along with channel attribution and device type information. This visibility helps you understand LLM traffic patterns and optimize your documentation for AI consumption.
+This visibility helps you understand LLM traffic patterns and optimize your documentation for AI consumption.
 
-## Technical Details
 
-The llms.txt generation system handles complex schema types including self-referencing schemas, preventing infinite recursion during content generation. For large documentation sites, the system uses optimized timeouts to ensure reliable content delivery without timing out.
-
-For REST API endpoints, the markdown generation includes OpenAPI specifications in YAML format, making it easy for LLMs to understand your API structure. The same markdown generation logic is used for both serving llms.txt files and indexing content for search, ensuring consistency across your documentation platform.
-
-## View in Action
+## View in action
 
 Check out the llms.txt files for this site:
 - `https://buildwithfern.com/learn/llms.txt`