Agent Memory

VoltAgent's Memory class stores conversation history and enables agents to maintain context across interactions. Supports persistent storage, semantic search, and working memory.

Storage Providers

Provider	Package	Persistence	Use Case
InMemory	@voltagent/core	None (RAM only)	Development, testing
Managed Memory	@voltagent/voltagent-memory	VoltOps-hosted	Production-ready, zero-setup
LibSQL	@voltagent/libsql	Local SQLite or remote	Self-hosted, edge deployments
Postgres	@voltagent/postgres	Self-hosted Postgres	Existing Postgres infrastructure
Supabase	@voltagent/supabase	Supabase	Supabase-based applications

Core Features

• Conversation Storage - Messages stored per userId and conversationId
• Semantic Search - Retrieve past messages by similarity (requires embedding + vector adapters)
• Working Memory - Compact context storage (Markdown template, JSON schema, or free-form)
• Workflow State - Suspendable workflow checkpoint storage

Quick Start

import { Agent, Memory } from "@voltagent/core";
import { ManagedMemoryAdapter } from "@voltagent/voltagent-memory";
import { openai } from "@ai-sdk/openai";

const memory = new Memory({
  storage: new ManagedMemoryAdapter({
    databaseName: "my-app-memory",
  }),
});

const agent = new Agent({
  name: "Assistant",
  model: openai("gpt-4o-mini"),
  memory,
});

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

// Agent remembers context
await agent.generateText("What's my name?", {
  userId: "user-123",
  conversationId: "chat-001",
});

Complete Documentation

For detailed configuration, provider setup, and advanced features:

• Memory Overview - Full memory system documentation
• Managed Memory - Production-ready hosted storage
• Semantic Search - Vector-based message retrieval
• Working Memory - Compact context management
• Storage Adapters - Provider-specific guides