Caching & Headers

How responses are cached at the edge and what every header means.

Cache strategy

Transcript responses are cached at the edge for 7 days by default. Subsequent requests for the same video / language / format combination return in <50ms median. Caching is transparent — request and response shapes are identical. Cached HITs still cost 1 credit.

X-Cache-Status

The X-Cache-Status header tells you where the response came from:

HIT — Fully served from cache. Median 49ms.
PARTIAL-HIT — Some segments served from cache, others fetched fresh.
MISS — Fetched live from the upstream pipeline.

http

HTTP/1.1 200 OK
X-Cache-Status: HIT
X-RateLimit-Limit: 300
X-RateLimit-Remaining: 287
X-RateLimit-Reset: 1735689600
Content-Type: application/json

Cache invalidation

Currently videos are cached for 7 days. We invalidate proactively when YouTube reports the video has been updated, but you should generally treat the cache as eventually consistent within the TTL.