Making the docs AI-friendly

AI and the era of "vibe coding" with LLM tools like Cursor and Copilot are changing the way we write code.

"I love reading docs" said no one ever. To help developers like you, we're adopting LLMs.txt, a new approach to making documentation more accessible and AI-friendly.

The gist of LLMs.txt

LLMs.txt is quickly becoming a standard way to structure documentation in a way that Large Language Models (LLMs) can understand. Learn about LLMs.txt

With the LLMs.txt format in place, you can:

• Get more accurate answers from Cursor, Copilot, and other LLM tools.
• Migrate to v3 from v2 faster since most LLMs are currently trained on v2 code.

How to use LLMs.txt

:::steps

Choose the txt file

Based on the available context window of the LLM you're using, choose the txt file that fits your needs.

• llms.txt: The main LLMs.txt file
• llms-full.txt: The complete documentation for Chakra UI v3
• llms-components.txt: The complete documentation for all components
• llms-component/[component].txt: Documentation for a specific component
• llms-styling.txt: Only styling documentation
• llms-theming.txt: Only theming documentation
• llms-migration.txt: Migration guide for v2 to v3

Add it to your LLM

You can upload the txt file as a knowledge base in Cursor, add it to your ChatGPT sidebar, or use it with NotebookLM.

Here's a quick demo of adding it to Cursor

<Video src="/videos/llms-txt-demo.mp4" />

Enjoy!

You can now ask the LLM questions about the specific components and get accurate answers.

:::

(Pro Tip) Migration

If you're migrating from v2 to v3, you can load up the llms-migration.txt file to get help with migrating your Chakra project much faster.

Conclusion

We hope this helps you get the most out of our docs and makes your coding experience even better.

Keep shipping!