langgraph-checkpoint-redis#
Redis implementations for LangGraph — checkpoint savers, stores with vector search, and agent middleware.
Understand the architecture, key patterns, and design decisions behind Redis-backed checkpointing, stores, and middleware.
Step-by-step guides for installation, configuration, TTL management, middleware setup, and enterprise deployment.
23 Jupyter notebooks covering checkpoints, human-in-the-loop, memory, middleware, and ReAct agents.
Complete API documentation for all public classes — savers, stores, middleware, and utilities.
Quick Start#
Install the package:
pip install langgraph-checkpoint-redis
Use Redis as your LangGraph checkpoint saver:
from langgraph.checkpoint.redis import RedisSaver
with RedisSaver.from_conn_string("redis://localhost:6379") as checkpointer:
checkpointer.setup()
# Use with any LangGraph graph
graph = builder.compile(checkpointer=checkpointer)
config = {"configurable": {"thread_id": "my-thread"}}
result = graph.invoke({"messages": [("human", "Hello!")]}, config)
Features#
Checkpoint Savers — Full and shallow variants, sync and async, with automatic TTL
Stores — Key-value storage with optional vector search for semantic retrieval
Middleware — Semantic caching, tool result caching, conversation memory, and semantic routing
Cluster Support — Automatic detection and handling of Redis Cluster, Azure Managed Redis, and Redis Enterprise
TTL Management — Native Redis TTL with refresh-on-read and per-thread pinning