Skip to main content

Ctrl+K

Getting Started

Quick Start
Build Guide

Deployment

Mooncake Store Deployment & Tuning Guide
Kubernetes Deployment Guide
Mooncake x SGLang Integration
Mooncake x vLLM Integration
Mooncake x LMCache Integration
Mooncake x LMDeploy Integration

Performance

Mooncake Performance
SGLang Integration Performance
vLLM Integration Performance

Design Documents

Mooncake Architecture
Transfer Engine
TENT: Transfer Engine NEXT
Mooncake Store
Mooncake PG Design
Mooncake EP Design
Mooncake Conductor
Mooncake Reshard
P2P Store
Mooncake x SGLang HiCache System Design

API Reference

Python
C/C++ API Reference
HTTP
Rust

Troubleshooting

Error Code Explanation
Troubleshooting
Mooncake PG/EP Troubleshooting

Community

Governance

Repository
Suggest edit

.md

Mooncake Reshard

Mooncake Reshard#

mooncake-reshard plans address-free conversions between complete resource placements, then binds the selected logical regions to immutable runtime snapshots.

Resource Reshard Manifest Contract
Model Weight Reshard Planner And Runtime Binding
Model Weight Store Upload Planning
Weight Management Architecture

previous

KV Event Subscriber Guide

next

Resource Reshard Manifest Contract

By the Mooncake Team

© Copyright 2026, Mooncake Team.