llms.txt

https://llmstxt.org/

brief background information, guidance, and links to detailed markdown files.

• https://langchain-ai.github.io/langgraph/llms.txt

Cursor + Windsurf native support for llms.txt

https://x.com/jeremyphoward/status/1902109312216129905

Cursor: @docs → Index

+ reflect on the urls in llms.txt file provided here
+ reflect on the input question 
+ directly read any urls in llms.txt that are relevant to the question 
+ use this to answer the question

question:
what are types of memory in LangGraph?

Windsurf: @docs → Search over domain

Approaches

• Context stuffing
  • Pro: Simple
  • Con: Costly, less efficient (for large docs), “lost in the middle”
• Vectorstore
  • Pro: Cheap (embeddings), scalable, fast
  • Con: Tricky to configure (split size, K), up-front work to set up / host vectorstore
• llms.txt

  • Pro: Simple, easy to audit, cheap

  • Con: Up-front work to build llms.txt, higher latency

    

MCP Server for llms.txt

https://www.youtube.com/watch?v=CDjjaTALI68

https://github.com/langchain-ai/mcpdoc

MCP server

https://github.com/langchain-ai/mcpdoc