Metadata-Version: 2.4
Name: openlake-vllm
Version: 0.8.1
Summary: OpenLake high performance KV cache client and server for vLLM.
License: Apache-2.0
Requires-Python: >=3.10
Description-Content-Type: text/markdown; charset=UTF-8; variant=GFM

# OpenLake for vLLM

`openlake-vllm` packages the OpenLake KV-cache client, the `openlaked` daemon,
the OpenLake vLLM connector, and the OpenLake control-plane UI for Linux.

Version 0.8.1 provides the non-RDMA local/H2 transport as a
`cp310-abi3-manylinux_2_28_x86_64` wheel. It supports regular CPython 3.10 and
newer on x86-64 Linux. Install vLLM separately for the connector runtime.

The RDMA/UCX build, ExANS CUDA codec, source distribution, and aarch64 wheel
are not included in this package release.

Project source and documentation are available at
[github.com/openlake-project/openlake](https://github.com/openlake-project/openlake).

