Search Packages

lmcache

Supercharge Your LLM with the Fastest KV Cache Layer

quantcpp

LLM inference with 7x longer context. Pure C, zero dependencies. Lossless KV cache compression + single-header library.

turboquant-vllm

TurboQuant KV cache compression plugin for vLLM — asymmetric K/V, 8 models validated, consumer GPUs

llm-autotune

Zero-config local LLM optimization for Ollama, LM Studio, and Apple Silicon MLX. Reduces TTFT by 40%, wall time for local agents by 46%, and RAM usage by 3x.

turboquant

First open-source TurboQuant KV cache compression for LLM inference. Drop-in for HuggingFace. pip install turboquant.

fade-kv

Frequency-Adaptive Decay Encoding: Attention-aware tiered KV cache compression for LLM inference.

turboquant-mlx-full

Extreme weight and KV cache compression for LLMs on Apple Silicon (MLX implementation of Google's TurboQuant)

avp

Python SDK for Agent Vector Protocol – transfer KV-cache between LLM agents instead of text

tqai

TurboQuant KV cache compression for local LLM inference

cappr

Completion After Prompt Probability. Make your LLM make a choice

prismkv

3-D Stacked-Plane KV Cache Quantizer — defensive prior art publication

spectral-kv

Up to 28x KV cache compression for LLMs via spectral SVD projection. Practically lossless on modern architectures.

nexusquant-kv

Training-free KV cache compression for LLMs. 10-33x compression via E8 lattice quantization + attention-aware token eviction. One line of code.

kvfleet

Production-grade, KV-cache-aware intelligent routing for self-hosted and hybrid LLM fleets.

kakeyalattice

Discrete Kakeya cover for LLM KV cache: D4/E8 nested-lattice quantisation realising a Kakeya-style tube-cover over the direction sphere. 2.4x-2.8x compression at <1% perplexity loss on Qwen3, Llama-3, DeepSeek, GLM-4, Gemma. Drop-in transformers.DynamicCache. pip install kakeyalattice.

adaptq

High-performance CPU KV-cache quantization engine for LLM inference (~10× speedup, 4× memory reduction) with Python & PyTorch support.

kvat

Automatic KV-Cache optimization for HuggingFace Transformers. Find the optimal cache strategy, attention backend, and dtype for your LLM inference workload.

lmcache-cli

A LLM serving engine extension to reduce TTFT and increase throughput, especially under long-context scenarios.

fused-turboquant

Fused Triton encode/decode kernels for TurboQuant KV cache compression, powered by Randomized Hadamard Transform.

langchain-turboquant

TurboQuant vector store for LangChain — 6x memory reduction with training-free quantization

turbokv

First open-source implementation of TurboQuant (arXiv 2504.19874) — 4-7x LLM KV cache compression

phase-slip-sampler

Phase-Slip is a stochastic intervention architecture that operates on the Key-Value Cache of the model. Phase-Slip gently rotates the semantic vectors of the context window, asking the model: "How would you finish this sentence if you looked at it from a slightly different perspective?"

turboquant-impl

First open-source implementation of TurboQuant (arXiv 2504.19874) — 4-7x LLM KV cache compression

kvcache-bench

Benchmark every KV cache compression method on your GPU. One command, real numbers. Supports Ollama + llama.cpp.