GitHub doesn't show folder sizes anywhere in its interface — not in the file listing, not in the API docs' most obvious endpoints, not even as an option in settings. Here's the fastest way to actually get that number, and why the "clone it and check" approach most people default to is more work than it needs to be.
Ask most developers how to check a GitHub folder's size and they'll say "clone the repo and run du -sh foldername." That works, but it means downloading the entire repository — every branch reference, every file, potentially hundreds of megabytes — just to answer a question about one directory. For a quick curiosity check, that's a disproportionate amount of work, bandwidth, and disk space for the answer you're actually after.
There's a lighter-weight path: GitHub's own API already knows every file's size, because it has to in order to serve the repository at all. The trick is that this data isn't rendered anywhere in the normal github.com interface — it exists, but you have to go get it deliberately.
Because RepoKit reads size metadata directly from GitHub's Git Trees API rather than downloading file contents, you get an answer without transferring the files themselves at all — the size check itself uses a fraction of the bandwidth that actually downloading the folder would.
If you genuinely only need this one time, GitHub's API is public and can be queried directly with a tool like curl against the Git Trees endpoint, summing the size field across matching paths yourself. It works, but it requires knowing the exact API shape, handling pagination for large trees, and doing the arithmetic by hand — which is exactly the manual work RepoKit exists to skip. For anyone checking folder sizes more than once, the extension pays for itself in the first use.
Install RepoKit once and size badges appear on every repo you visit from then on.
Add RepoKit to Chrome — it's free
Repo