Install¶

Works with both pip and uv — pick whichever you use.

pipuv

pip install memory-reuse

uv add memory-reuse

The core has zero required dependencies. Optional features are opt-in via extras.

Optional extras¶

Extra	What it adds	Install
`redis`	Redis backend support	`pip install "memory-reuse[redis]"`
`litellm`	LiteLLM cached wrappers	`pip install "memory-reuse[litellm]"`
`semantic`	Semantic cache with API embeddings (OpenAI / LiteLLM) — no torch	`pip install "memory-reuse[semantic]"`
`semantic-local`	Semantic cache with local embeddings (sentence-transformers, pulls in torch)	`pip install "memory-reuse[semantic-local]"`
`all`	Everything above	`pip install "memory-reuse[all]"`

Two commands cover every embedding provider

The semantic extra bundles the small openai and litellm clients (plus numpy) and installs no torch — enough for the OpenAI and LiteLLM embedding providers. Only the local provider needs semantic-local.

Local embeddings and PyTorch¶

sentence-transformers (the local provider) depends on PyTorch. On a CPU-only machine, install the CPU torch wheel first to avoid a ~2 GB GPU/CUDA download — the CPU build is ~200 MB:

pip install torch --index-url https://download.pytorch.org/whl/cpu
pip install "memory-reuse[semantic-local]"

With a GPU you can skip the first line. Either way, the model weights (e.g. all-MiniLM-L6-v2, ~90 MB) download from Hugging Face on first use and are then cached on disk for offline reuse.

Quieter / fully offline runs

memory-reuse silences the Hugging Face log chatter and the per-embedding progress bar. Once the model is cached, you can also skip Hugging Face's HTTP cache-validation checks by exporting HF_HUB_OFFLINE=1 (and TRANSFORMERS_OFFLINE=1) in your own application — never inside a shared library, and only after the first (downloading) run.