Skip to main content

Agent Memory

Memory enables agents to remember past interactions and maintain conversation context. This guide shows how to configure memory for your agents.

Default Behavior​

Agents automatically use in-memory storage by default, storing conversations in application memory:

import { Agent } from "@voltagent/core";
import { openai } from "@ai-sdk/openai";

const agent = new Agent({
  name: "Assistant",
  instructions: "You are a helpful assistant",
  model: openai("gpt-4o"),
  // Memory is automatically enabled with in-memory storage
});

Using Memory​

To maintain conversation context, provide userId and conversationId:

// First message
const response1 = await agent.generateText("My name is Sarah", {
  userId: "user-123",
  conversationId: "chat-001",
});

// Follow-up message - agent remembers the name
const response2 = await agent.generateText("What's my name?", {
  userId: "user-123",
  conversationId: "chat-001", // Same conversation ID
});
console.log(response2.text); // "Your name is Sarah"

Persistent Storage​

For conversations that survive application restarts, configure Memory with a persistent adapter such as LibSQL:

import { Agent, Memory } from "@voltagent/core";
import { LibSQLMemoryAdapter } from "@voltagent/libsql"; // npm install @voltagent/libsql
import { openai } from "@ai-sdk/openai";

const agent = new Agent({
  name: "Persistent Assistant",
  instructions: "You are a helpful assistant",
  model: openai("gpt-4o"),
  memory: new Memory({
    storage: new LibSQLMemoryAdapter({ url: "file:./.voltagent/memory.db" }),
  }),
});

Disabling Memory​

For stateless agents that don't need conversation history:

const agent = new Agent({
  name: "Stateless Assistant",
  instructions: "You provide one-off responses",
  model: openai("gpt-4o"),
  memory: false, // Disable memory completely
});

Available Memory Providers​

Learn More​

For detailed information about memory configuration, providers, and advanced usage:

Table of Contents