Appearance

Running locally ​

Bring the whole stack up on your machine. These are developer instructions for the RAGBot monorepo — end users don't need any of this.

Prerequisites ​

  • Node.js >= 24 with Corepack (bundled with Node 24+)
  • PostgreSQL with the pgvector extension (for the API)
  • Redis (any Redis-compatible service) for the background job queue
  • API keys for the external services the pipeline uses: OpenRouter (embeddings + chat), Brevo (email), S3/R2 (file storage), LlamaIndex (PDF/Word parsing), Firecrawl (URL scraping)

Get the code ​

bash
git clone https://github.com/dnwandana/ragbot.git
cd ragbot

Install ​

bash
corepack pnpm install

Configure the API ​

bash
cp apps/api/.env.example apps/api/.env

Fill in the required variables — database and Redis URLs, two distinct JWT secrets (≥32 chars each), and the external API keys above. The API validates its environment at startup and exits if anything required is missing, so a misconfigured .env fails fast with a clear message.

The browser app reads one variable, VITE_API_BASE_URL (defaults to http://localhost:3000/api).

Local cookies need development mode

Set NODE_ENV=development locally. The API only marks auth cookies Secure in production, and browsers reject Secure cookies over plain HTTP — so a production config won't let you log in over http://localhost.

Run the database migrations ​

Migrations don't run automatically:

bash
corepack pnpm --filter @ragbot/api migrate

There are also seeds for the 31 permissions and a couple of test users.

Start the apps ​

bash
corepack pnpm dev          # all apps via Turborepo
# — or individually —
corepack pnpm dev:api      # API on :3000
corepack pnpm dev:app      # SPA on :8080
corepack pnpm dev:web      # marketing site on :4321
corepack pnpm dev:docs     # these docs on :4173

Open the app at http://localhost:8080.

Other useful commands ​

bash
corepack pnpm build        # build every app
corepack pnpm lint         # lint all
corepack pnpm test:api     # API test suite (Vitest + Supertest against real PostgreSQL)

Running in containers ​

To run the whole stack in Docker — locally or in production — see Deployment.