runtime-cache

You can cache API responses across serverless functions using an ephemeral per-region key-value cache. It leverages the @vercel/functions API surface to store and retrieve computed data regionally across Functions, Routing Middleware, and Builds within your deployment.

Tag-based cache invalidation groups cached keys by specific tags, allowing granular cache coherence and bulk purges. When data changes, you invalidate by tag to ensure stale serverless cache entries are removed without clearing the entire regional store.

Yes, the Vercel runtime cache provides per-region isolation for cross-function state. This ephemeral key-value store ensures regional data sharing without cross-region latency, keeping cached data close to your serverless execution environment.

Yes, you can use TTL-based caching and namespace options for cache management in serverless workflows. These features, alongside optional key hashing, provide control over cache lifespans and observability across your project deployment environment.

An ephemeral regional cache is temporary and does not persist data permanently across deployments. It is limited to regional isolation and TTL constraints, meaning cached data expires and should not be relied upon for durable, long-term cross-function state storage.