It's not the core of koda-validate, and yeah lots of libraries have a similar capacity. Feedback I'd be interested in is if there are gaps.

In general the value prop of koda-validate is that it turns validation into typesafe building blocks, which makes validators very re-usable -- and flexible. Some other notable differences from pydantic are that it doesn't `raise` on validation errors, you don't need a typing plugin, and it's fully asyncio-compatible.