Skip to main content
The Rowboat web platform is a Next.js-based dashboard that provides a comprehensive interface for managing AI agents, projects, and integrations.

​
Architecture

The web platform is built on a modern stack designed for scalability and performance:
  • Framework: Next.js 15 with React 19
  • Runtime: Node.js 18+ (Alpine Linux in containers)
  • Language: TypeScript 5 with ES2022 target
  • Styling: TailwindCSS with Radix UI components
  • AI SDK: Vercel AI SDK with multi-provider support

​
Core Services

The platform relies on three key infrastructure services:

​
MongoDB

Primary database for storing:
  • User accounts and authentication data
  • Project configurations
  • Agent definitions and state
  • Chat history and conversations
  • Job queue metadata
Default Connection: mongodb://mongo:27017/rowboat

​
Redis

In-memory data store used for:
  • Session management
  • Job queue processing (BullMQ)
  • Caching frequently accessed data
  • Rate limiting counters
Default Connection: redis://redis:6379

​
Qdrant

Vector database for RAG (Retrieval Augmented Generation):
  • Document embeddings storage
  • Semantic search capabilities
  • File upload content indexing
  • Web scraping data storage
Default Connection: http://qdrant:6333
Qdrant is optional and only required if you enable RAG features (USE_RAG=true).

​
Platform Features

​
AI Agent Management

Create and manage AI agents with:
  • Custom instructions and personas
  • Tool integrations (Composio, Klavis)
  • Multi-provider LLM support (OpenAI, Anthropic, Google, custom)
  • Agent-to-agent communication

​
RAG Capabilities

When enabled, the platform supports:
  • File Uploads: Process and embed documents (PDF, TXT, etc.)
  • URL Scraping: Extract and index web content using Firecrawl
  • S3 Storage: Optional cloud storage for uploaded files
  • Semantic Search: Query documents using natural language

​
Authentication

Optional Auth0 integration for:
  • User registration and login
  • OAuth social providers
  • Multi-tenant isolation
  • Role-based access control

​
Background Workers

The platform includes specialized worker processes: RAG Worker (rag-worker)
  • Processes file uploads
  • Generates embeddings
  • Indexes content in Qdrant
  • Handles web scraping jobs
Jobs Worker (jobs-worker)
  • Executes scheduled tasks
  • Processes background jobs
  • Manages agent workflows
  • Handles Composio integrations

​
Technology Stack

ComponentTechnology
FrontendReact 19, Next.js 15
UI LibraryHeroUI, Radix UI
State ManagementReact hooks, Immer
AI SDKVercel AI SDK
DatabaseMongoDB 6.8+
Cache/QueueRedis (ioredis)
Vector DBQdrant
File StorageLocal filesystem or AWS S3
AuthenticationAuth0 (optional)
API IntegrationComposio, Klavis
Build Toolesbuild, Turbopack

​
Development vs Production

​
Development Mode

npm run dev
Features:
  • Hot module replacement
  • Turbopack for fast builds
  • Source maps enabled
  • Runs on http://localhost:3000

​
Production Build

npm run build
npm start
Features:
  • Optimized bundle size
  • Standalone output for Docker
  • Static asset optimization
  • Server-side rendering
The production build requires all environment variables to be set at build time. Make sure your .env file is complete before building.

​
Next Steps

Deployment

Learn how to deploy the web platform using Docker

Configuration

Configure environment variables and features

Build docs developers (and LLMs) love

Get started for freeTalk to us