Skip to main content

In-Memory Storage

InMemoryStorageAdapter stores conversation history in application memory. Data is lost when the application restarts.

Default Behavior​

Agents use in-memory storage by default when no memory option is provided:

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

// Uses InMemoryStorageAdapter automatically
const agent = new Agent({
  name: "Assistant",
  instructions: "Help users with questions.",
  model: openai("gpt-4o-mini"),
});

Explicit Configuration​

Configure storage limits explicitly:

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

const memory = new Memory({
  storage: new InMemoryStorageAdapter({
    storageLimit: 100, // max messages per userId/conversationId (default: 100)
  }),
});

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

Features​

Conversation Storage​

  • Messages stored per userId and conversationId
  • Oldest messages pruned when storageLimit exceeded
  • All StorageAdapter methods supported

Working Memory​

Supports both conversation and user-scoped working memory:

const memory = new Memory({
  storage: new InMemoryStorageAdapter(),
  workingMemory: {
    enabled: true,
    scope: "conversation", // or "user"
  },
});

See Working Memory for configuration details.

Semantic Search (Development)​

Combine with InMemoryVectorAdapter for semantic search during development:

import {
  Memory,
  AiSdkEmbeddingAdapter,
  InMemoryVectorAdapter,
  InMemoryStorageAdapter,
} from "@voltagent/core";
import { openai } from "@ai-sdk/openai";

const memory = new Memory({
  storage: new InMemoryStorageAdapter(),
  embedding: new AiSdkEmbeddingAdapter(openai.embedding("text-embedding-3-small")),
  vector: new InMemoryVectorAdapter(),
});

Both storage and vectors are lost on restart. For persistent vectors, use LibSQLVectorAdapter.

Use Cases​

Development & Testing​

Test agent logic without database setup:

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

const testAgent = new Agent({
  name: "Test Assistant",
  model: openai("gpt-4o-mini"),
  memory: new Memory({
    storage: new InMemoryStorageAdapter({ storageLimit: 50 }),
  }),
});

// Test conversations without persistence
await testAgent.generateText("Test message", {
  userId: "test-user",
  conversationId: "test-conversation",
});

Stateless Deployments​

Serverless functions or ephemeral containers where persistence isn't needed:

// Cloud function handler
export async function handler(event) {
  const agent = new Agent({
    name: "Serverless Assistant",
    model: openai("gpt-4o-mini"),
    // Default in-memory storage
  });

  return await agent.generateText(event.message, {
    userId: event.userId,
    conversationId: event.sessionId,
  });
}

Demos & Examples​

Quick prototypes without infrastructure dependencies.

Limitations​

  • No persistence - All data lost on restart
  • Memory usage - Large message counts consume application memory
  • Not for production - Use persistent adapters for production applications

Learn More​

Table of Contents