Hey, I’m wondering can we ask cursor AI to go through and use multiple codebases when answering questions? e.g., if my current project is using / importing functions from two other packages/codebases and these two packages do not have documentaions, is there a way to ask cursor to check and use these two additional codebases when phrasing the response? I see there’s an @codebase function, but I’m not sure whether it is for this function, nor about how to use it.
1 Like
I’d like know this too.
I have multiple projects in a local repository folder that it could be really beneficial for cursor to be able to scan and index them.
3 Likes
I’m also interested in this. My current use-case is that I have multiple apps to deploy on a single server and I’d like some AI recommendations on best practices for server configuration, but I want the AI to have the whole context of what needs to be deployed
Right now there’s no direct way to have chat reference multiple codebases at once, but there’s a workaround you can try - you can use chat to summarize the implementation details of one codebase, then paste that summary as context in a new chat when looking at the other codebase
For your server config case, you could:
- Open each app’s codebase separately in Cursor
- Use @codebase to get a summary of each app’s implementation and requirements
- Create a new chat with your server config files open
- Paste in the summaries from step 2 as context
- Ask for configuration recommendations with all that context available
Not ideal but should help get you the context you need until we add proper multi-codebase support