This example shows how to use Workflow to make RAG agents more durable and reliable. It is based on the Retrieval-Augmented Generation (RAG) guide.

• An API key from Vercel AI Gateway
• A PostgreSQL database

1. Clone the repository and install dependencies:

   git clone https://github.com/vercel/workflow-examples
   cd workflow-examples/rag-agent
   pnpm install

2. Create a .env.local file in examples/rag-agent/:

   touch .env.local

3. Add your Vercel AI Gateway API key and database URL to the .env.local file:

   AI_GATEWAY_API_KEY=your_api_key_here
   DATABASE_URL=your_database_url_here

4. Run the migrations:

   pnpm db:migrate

5. Start the development server:

   pnpm dev

6. Open http://localhost:3000 to see the app

This project uses the following stack:

• Next.js 15 (App Router)
• Vercel AI SDK
• Vercel AI Gateway
• Drizzle ORM
• Postgres with pgvector
• shadcn-ui and TailwindCSS for styling