reflex-cache 0.2.1

Episodic memory and high-speed semantic cache for LLM responses
Documentation

reflex-cache

reflex-cache is the core Reflex library: tiered cache, storage, embeddings/reranking, and vector DB integration.

This crate is published to crates.io as reflex-cache, but exposes its library as reflex (so downstream code can use reflex::...).

Docs: https://docs.rs/reflex-cache

Use As A Dependency

[dependencies]
reflex = { package = "reflex-cache", version = "x.x.x" }

Quick Start

From this repo:

cargo run -p reflex-cache --example basic_lookup --features mock

What’s Inside

  • cache: tiered cache orchestration (L1 exact + L2 semantic)
  • storage: rkyv-backed storage + mmap/NVMe loaders
  • vectordb: Qdrant client + binary quantization helpers (and mocks behind mock)
  • embedding: embedder + reranker wiring
  • scoring: L3 verification (cross-encoder)
  • config: env-backed configuration types used by the server

Features

  • metal: Apple Silicon acceleration (passes through to model deps)
  • cuda: NVIDIA acceleration (passes through to model deps)
  • mock: enables mock backends for tests/examples

Build / Test

From repo root:

cargo test -p reflex-cache --features mock
cargo doc -p reflex-cache --no-deps