This example demonstrates how to use LanceDB as a vector database/retriever within a VoltAgent application.

• Local & Serverless: Uses LanceDB which runs embedded locally—no Docker or API keys required (unless using LanceDB Cloud).
• Multimodal Ready: LanceDB is optimized for multimodal data (text, images, video), making this a future-proof foundation.
• Automatic Initialization: Automatically creates the knowledge base table and populates it with sample data on first run.
• Semantic Search: Uses OpenAI embeddings to retrieve relevant documents based on user queries.
• Two Agent Patterns:
  1. Assistant with Retriever: Automatically uses retrieved context for every message.
  2. Assistant with Tools: Autonomously decides when to use the retrieval tool.

• Node.js 20+
• OpenAI API Key (for embeddings and LLM)

1. Install dependencies:

   npm install

2. Configure Environment: Copy .env.example to .env and add your OpenAI API Key:

   cp .env.example .env

   Edit .env:

   OPENAI_API_KEY=sk-...

3. Run the Agent:

   npm run dev

• The database is stored locally in .voltagent/lancedb.
• On startup, src/retriever/index.ts checks if the table exists.
• If not, it creates it and indexes the sample documents defined in the code.
• Agents can then query this local database with low latency.