This isn't really a feature request, but I'm not sure where else to put something like this for the Kagi team to see.
uv is a popular OSS project in the Python ecosystem. They are moving really fast and the popular LLMs struggle to give accurate answers. I think that's due to stale training data but also possibly that "uv" is too short and causes problems when indexing/predicting or whatever the LLMs do with it.
There are also a ton of good (but also very stale) info in the GH repo that could probably answer a lot of questions if the newer info was prioritized over the old.
If this is something Kagi Sidekick could help with, maybe someone from Kagi wants to jump into the issue I created to discuss the possibility: https://github.com/astral-sh/uv/issues/13711#issuecomment-2929829891
I'd love to see the uv docs site have a search/chat combination that would give them more accurate replies based on the most current uv docs and best practices and also help point them to issues in the GitHub repo when appropriate.
