Learn how to use LangChain Memory to retain and recall information across interactions — a vital feature for chatbots, virtual agents, and intelligent assistants.

---

✨ Why Memory Matters in LLM Apps

Large Language Models (LLMs) are stateless by default. Without memory, your assistant forgets everything after each prompt. LangChain solves this with powerful memory modules:

• 🗣️ Multi-turn conversations
• 🧠 Personalized interactions
• 🔄 Efficient context retention

---

🧩 Types of Memory in LangChain

Memory Type	Purpose
BufferMemory	Stores full conversation history as-is
ConversationSummaryMemory	Summarizes past messages using an LLM for brevity
EntityMemory	Tracks entities like names, topics, objects
VectorStoreRetrieverMemory	Retrieves past data semantically using vector similarity

Each memory type fits different needs — from long-term factual recall to short-term conversational continuity.

---

🛠️ Hands-On Example: BufferMemory

import { ConversationChain } from "langchain/chains";
import { ChatOpenAI } from "langchain/chat_models/openai";
import { BufferMemory } from "langchain/memory";

const memory = new BufferMemory();
const chain = new ConversationChain({ llm: new ChatOpenAI(), memory });

await chain.call({ input: "Hi, I'm Bhanu" });
await chain.call({ input: "What’s my name?" });

Output: “Your name is Bhanu.”

The assistant remembers context across turns using memory.

When to Use Each Memory

• Use BufferMemory for raw dialogue retention
• Use ConversationSummaryMemory when saving tokens matters
• Use EntityMemory to build topic-aware agents
• Use VectorStoreRetrieverMemory in document Q&A bots

Memory can be attached to Chains, Agents, or even Tools.

Related Tutorials

📘 LangChain Mastery Series

🔗 Hands-On: LangChain Prompt Templates