> ## Documentation Index
> Fetch the complete documentation index at: https://docs.tensormesh.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Cache Sizing

> Paste a chronological trace of token-id requests and simulate the cache hit rate you would get at any L1/L2 capacity, in tokens and bytes.

Trace — ordered list of token-id requests

JSON array, chronological. Each request is an array of token ids, e.g.
`[[1,2,3,4],[1,2,3,9]]`. Requests sharing a leading prefix share cached KV.

Bytes per token

Get it for your model from the
[LMCache KV Cache Size Calculator](https://docs.lmcache.ai/getting_started/kv_cache_calculator.html)
. Set it to also see capacities in bytes.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.