Does Supplying Documentation to Cursor Improve Answer Quality?

Discussions
1

I’ve been wondering how important it is to supply documentation to Cursor.

For example, I use Cursor primarily for iOS development. Should I upload all relevant documentation related to SwiftUI, UIKit, and Swift?

cursor
cursor1527×694 115 KB

I’m asking because, when coding without Cursor and using something like Gemini Pro 2.5, I still get high-quality responses, even without explicitly supplying any documentation.

So, I’m curious: when I use Cursor and provide iOS-related documentation, does it actually help improve the quality and accuracy of the answers?

Thanks.

2

AI’s are at least 6months to 1 year out of date as it takes time to collect data, convert it, feed it as training, extract evaluation data,…

Providing docs with the right version URL helps AI when you reference docs to use indexed up to date info. Cursor indexes it so the AI can fetch the matching content it needs and not whole docu which would pollute context.

Some models are post training tuned with certain coding tasks, programming languages and best practices for codeing. Which applies to Gemini too. However indexed up to date docu is preferrable as the model wont change after today for a long while.

You dont need to add docu to context all the time. It helps to do it when planning tasks or when making detailed decisions / changes on code that are new so the AI doesnt have project internal context to refer to.