Home Page | Documentation | Examples | Discord | Blog

This example shows how to enable semantic memory using embeddings and a vector database. Messages are embedded on save, stored in LibSQL via LibSQLVectorAdapter, and retrieved with semantic search when the agent answers.

npm create voltagent-app@latest -- --example with-vector-search

• OpenAI embeddings via AI SDK (text-embedding-3-small)
• Vector DB via LibSQLVectorAdapter
• Automatic semantic recall in Agent when vector support exists
• Optional raw vector search API (memory.searchSimilar)

import { Agent, Memory, VoltAgent } from "@voltagent/core";
import { LibSQLMemoryAdapter, LibSQLVectorAdapter } from "@voltagent/libsql";
import { honoServer } from "@voltagent/server-hono";

const memory = new Memory({
  storage: new LibSQLMemoryAdapter(),
  embedding: "openai/text-embedding-3-small",
  vector: new LibSQLVectorAdapter(),
});

const agent = new Agent({ name: "Semantic Memory Agent", model: "openai/gpt-4o-mini", memory });
new VoltAgent({ agents: { agent }, server: honoServer({ port: 3142 }) });

1. Copy .env.example to .env and set OPENAI_API_KEY.
2. Install deps and start:

pnpm i
pnpm dev

To observe semantic recall, seed a conversation with prior messages (see commented demo in src/index.ts), then ask a related question. The agent will merge relevant past messages using semantic search.