Improve Documentation UI for Better Readability and Navigation
Hi there 👋,
I was exploring the documentation of OpenHands and noticed that the current UI could be more user-friendly in terms of readability and navigation.
Is there any particular reason behind the current design of the docs UI?
I’d love to help improve it! If you're open to suggestions, I recommend a cleaner and more modern layout—something similar to DocumentX. This is my repo and it is open source. I would love to add custom components too. I am more than happy to help with the migration. I believe this could enhance the user experience significantly.
I'd be happy to contribute and help rebuild the docs UI of OpenHands. Let me know your thoughts!
Thanks 😊
Hi there. Thanks very much for your interest! I believe there was some talks about possibly using something else for the documentation. If it's going to be a big change, you will have to get some approval from some folks. I'll tag them in case they are already thinking of going another route. CC @rbren @xingyaoww @neubig
Hi there, we just finished migrating our doc site via mintlify today, so hopefully it is cleaner, easier to navigate, and can resolve this issue: https://docs.all-hands.dev/! Feel free to re-open if you think there are other ways to improve it 🙌
https://allhandsai.slack.com/archives/C07EM5CLPLZ/p1748639043727409
Hey @xingyaoww, thanks for the update! I’m asking purely out of curiosity—not to be critical or question your decision.
I noticed many open-source projects go with Mintlify, but since it’s neither open source nor free, I was wondering if it makes financial sense long-term for All Hands?
Also, now that All Hands will only be writing docs (not maintaining the site itself), will this limit UI customization?
Lastly, will there be any plans to integrate a chatbot on top of the documentation? Something like an in-house solution using chunking, embeddings, and tree summaries—maybe similar to what LlamaIndex or Kapa provide? I understand that mintlify has assistant service but will it not be more costly rather than just paying for embedding LLM API.
Appreciate any thoughts on this! 🙌