Theoritically this makes sense, however, the size returned by that API call is inaccurate. I'm getting 1.19MB as total size. The current implementation with this repo as test subject at least results in an accurate decompressed main branch download size. Neither implementation at the moments take into account other branches.

To replicate:

• Download the source code
• Extract the zip
• See folder size

@AminoffZ i think the truth lies somewhere in the middle the api call gives me 2.19 MB, the file size counter gives 1.31 MB, same time downloading and decompressing gives 1.30 MB and cloning gives 3.3MB, so i find it weird that one set of files on one specific branch that hasn't changed has so many changing size values based on the download/calculation method, what should we do?

I think the size return from API is counting all the storage space taken on github. It is not only counting the default branch file sizes.
I have a repo that the default branch only contains a readme file, but the size from API is much larger than the readme file size. The API value is almost equal to the .git folder size.