Comment by miketheman

PyPI responses are cached at 99% or higher, with less infrastructure to run.

Search is an unbounded context and does not lend itself to caching very well, as every search can contain anything