There's no official API or any documentation, so everything is reverse-engineered from the web UI. It's all rather straight-forward.
We use this to search and download IRD files for various purposes. One note though: we cache IRD files on-disk to limit the download requests.