Popular GitHub Repositories for Kvcache

Mooncake is the serving platform for Kimi, a leading LLM service provided by Moonshot AI.

R-KV: Redundancy-aware KV Cache Compression for Reasoning Models

kvcached: Elastic KV cache for dynamic GPU sharing and efficient multi-LLM inference.

PiKV: MoE KV Cache Management System [Efficient ML System]

A Unified Cross-Instance Scheduling Runtime for Heterogeneous Inference Engines (vLLM, SGLang, etc.), Supporting Multi-Level KV Cache Storage, Global Prefix Affinity Awareness, PD Disaggregation, and Advanced Intelligent Scheduling Features, etc.