LMCache Controller

LMCache Controller exposes a set of APIs for users and orchestrators to manage the KV cache.

Currently, the controller provides the following APIs:

• Lookup: Lookup the KV cache for a given list of tokens.

• Clear: Clear the KV caches.

• Pin: Persist or set the TTL of a KV cache.

• Move: Move the KV cache to a different location.

• Compress: Compress the KV cache.

• CheckFinish: Check whether a (non-blocking) control event has finished or not.