Examples

23 Jupyter notebooks demonstrating how to use langgraph-checkpoint-redis for persistent LangGraph workflows. Each notebook is self-contained and uses Redis as the backing store for checkpoints, memory, and middleware.

Checkpoints

Persist graph state across invocations using Redis-backed checkpoint savers, including cross-thread sharing, functional APIs, and subgraph patterns.

Notebook	Description
Cross-Thread Persistence	Cross-thread persistence with Redis stores
Cross-Thread Persistence (Functional)	Cross-thread persistence using the functional API
Functional Persistence	Functional persistence patterns
Subgraph Persistence	Subgraph checkpoint persistence
Managing State in Subgraphs	Managing state in subgraphs

Human-in-the-Loop

Interrupt graph execution for human review, edit state mid-run, and navigate checkpoint history with time travel.

Notebook	Description
Breakpoints	Adding breakpoints for human review
Dynamic Breakpoints	Dynamic breakpoints based on conditions
Edit Graph State	Editing graph state during execution
Review Tool Calls	Reviewing tool calls before execution
Review Tool Calls (OpenAI)	Reviewing tool calls with the OpenAI Responses API
Time Travel	Time travel to previous checkpoints
Wait for User Input	Waiting for user input during execution

Memory

Manage conversation history, summarize past messages, and perform semantic search over stored memories using RedisStore.

Notebook	Description
Summarize Conversation History	Summarizing conversation history
Delete Messages	Deleting messages from state
Manage Conversation History	Managing conversation history length
Semantic Search	Semantic search over stored memories

Middleware

Add caching, conversation memory, and semantic routing to LangGraph agents using the Redis middleware layer.

Notebook	Description
Semantic Cache	LLM response caching with semantic matching
Tool Result Caching	Tool result caching with metadata control
Conversation Memory	Semantic conversation history retrieval
Middleware Composition	Combining middleware with shared connections

ReAct Agents

Build ReAct agents with Redis-backed persistence, human-in-the-loop review, and long-term memory.

Notebook	Description
ReAct Agent with HITL	ReAct agent with human-in-the-loop
ReAct Agent with Message History	ReAct agent with message history management
ReAct Agent with Memory	ReAct agent with persistent memory