Doesn't solve rich data though as neatly as IndexedDB does.

Because IndexedDB uses structured clone, it can store javascript data as is, which includes storing correctly a number of interesting/useful types, including ArrayBuffers, Files/Blobs, CryptoKey instances (useful to avoid needing export flag on them), the cost of this is the performance hit compared to JSON.parse on a string.

Also the cache is less persistent than IndexedDB when storage pressure increases.