First memories

Extracting

Storing

Beyond extraction

Give a support assistant long-term memory with LangMem: extraction, updates, profiles, the LangGraph store, memory tools, background processing and summaries. No API key.

LangMem

LangMem overview

Install LangMem and set up API keys

Installation and setup

LangMem memory manager basics

create_memory_manager and first memories

How LangMem prompts the model

What LangMem asks the model

LangMem with a custom LangChain chat model

MemoryModel: a rule-based chat model

LangMem with Claude, GPT and other models

Real models: model strings and keys

LangMem semantic memory collections

Collections: many small memories

Update and delete memories in LangMem

Updating and deleting memories

LangMem profiles with Pydantic schemas

Schemas and profiles

LangGraph store for long-term memory

The LangGraph store: where memories live

Semantic memory search in LangGraph

Semantic search with an embedder

LangMem create_memory_store_manager

Store managers: extract and save in one step

LangMem memory tools for agents

Memory tools: manage and search

LangGraph agent with long-term memory

An agent that remembers across conversations

LangMem background memory processing

Background memory with ReflectionExecutor

LangMem summarize_messages for short-term memory

Summarizing long conversations

LangMem prompt optimizer

Prompt optimization: learning from feedback

LangMem assistant that remembers users

Support assistant that remembers customers

First memories quiz

What does manager.invoke({"messages": conversation}) return?