NetPad: MongoDB-Connected Form Builder Platform

Project Overview

1. Forms: Visual drag-and-drop form builder with 30+ field types, conditional logic, and multi-page support
2. Workflows: Automation engine with 25+ node types for post-submission processing and integrations
3. Data Management: Browse, query, and manage MongoDB data with visual tools and secure credential storage
4. AI/Conversational: Natural language data collection powered by RAG and MongoDB Atlas Vector Search

Key Features

Visual Form Builder

• WYSIWYG Editor: Drag-and-drop interface with live preview and keyboard shortcuts
• 30+ Field Types: Text, selection, date/time, rich input, files, signatures, geolocation, and more
• Multi-Page Forms: Step-based progression with configurable indicators (dots, numbers, progress bar, tabs)
• Conditional Logic: Show/hide fields based on other field values
• Computed Fields: Formula-based calculations with a powerful expression engine
• Lookup Fields: Cross-collection references with autocomplete
• Theming & Branding: Custom colors, logos, preset themes, and full CSS control
• Template Gallery: 25+ form templates and 11 workflow templates for rapid development

Form Types

Workflow Automation

• Visual Editor: Canvas-based ReactFlow-powered workflow design with node palette
• 25+ Node Types: Logic (conditional, switch, filter, loop), data transform, MongoDB operations, HTTP requests, email, Slack, Google Sheets
• Multiple Triggers: Form submission, webhook, schedule (CRON), manual, API
• Execution Engine: Queue-based with retry logic, exponential backoff, and comprehensive error handling
• Real-Time Monitoring: Execution status, progress tracking, and version history

Conversational Forms with RAG

• Natural Language Interaction: Users chat with an AI that guides data collection
• Topic Coverage Tracking: Ensures all required information is collected
• Automatic Data Extraction: Structured data with confidence scoring per field
• Knowledge-Guided (RAG): Ground AI responses in uploaded documents (PDF, DOCX, TXT)
• Source Citations: Every AI response includes traceable references to source documents
• Vector Search: MongoDB Atlas Vector Search for semantic document retrieval

Data Management

• Connection Vault: AES-256-GCM encrypted storage for MongoDB credentials
• Data Browser: Visual document viewing, editing, querying, and aggregation pipeline building
• Import/Export: CSV/JSON with schema inference and mapping
• Atlas Integration: Auto-provisioned M0 clusters, monitoring, and user management

Technical Implementation

Core Components

1. Form Processing Pipeline

2. RAG-Powered Conversational Forms

code-highlightAI: Hi! I'm here to help you submit an IT support ticket.
    What kind of issue are you experiencing?

User: My laptop won't turn on

AI: I'm sorry to hear that. When you say it won't turn on,
    does the screen stay completely black, or do you see any
    lights or error messages?

User: Completely black, no lights at all

AI: That sounds like a hardware issue. How urgent is this?
    Are you able to work from another device?

→ Extracted Data:
  {
    "issueCategory": "hardware",
    "urgencyLevel": "high",
    "description": "Laptop won't power on, no lights...",
    "confidence": 0.92
  }

3. Workflow Execution Engine

Database Schema

javascript code-highlight// Form Definition
{
  _id: ObjectId,
  name: String,
  slug: String,
  formType: 'entry' | 'search' | 'both' | 'conversational',
  fields: [{
    id: String,
    type: String,
    label: String,
    required: Boolean,
    validation: Object,
    conditionalLogic: Object
  }],
  theme: {
    primaryColor: String,
    headerStyle: String,
    preset: String
  },
  organizationId: ObjectId,
  projectId: ObjectId
}

// Workflow Definition
{
  _id: ObjectId,
  name: String,
  trigger: {
    type: 'form' | 'webhook' | 'schedule' | 'manual' | 'api',
    config: Object
  },
  nodes: [{
    id: String,
    type: String,
    position: { x: Number, y: Number },
    data: Object
  }],
  edges: [{
    source: String,
    target: String
  }],
  status: 'draft' | 'published'
}

// RAG Document
{
  _id: ObjectId,
  formId: ObjectId,
  title: String,
  chunks: [{
    content: String,
    embedding: [Number],  // 1536 dimensions
    metadata: { startIndex: Number, endIndex: Number }
  }],
  metadata: { fileType: String, uploadedAt: Date }
}

Platform Services

Organizations & Multi-Tenancy

code-highlightOrganization
├── Projects (dev, staging, prod)
│   ├── Applications
│   │   ├── Forms
│   │   ├── Workflows
│   │   ├── Connections
│   │   └── Releases (versioned snapshots)
│   ├── Forms (standalone)
│   └── Workflows (standalone)
├── Members (Owner, Admin, Member, Viewer)
├── Connection Vault
├── Templates
└── Billing/Subscription

Application Marketplace

• Browse: Discover published applications by category, tags, or search
• Import: One-click import into your projects
• Publish: Share your applications with the community
• npm Integration: Install applications directly from npm registry

AI Agents (15+ Available)

Challenges and Solutions

Challenge 1: Multi-Tenant Architecture

• Organization-scoped queries throughout the application
• Connection vault with encrypted credentials per organization
• Project-based resource grouping (dev, staging, prod)
• Fine-grained RBAC at organization, application, and form levels

Challenge 2: Real-Time Conversational Experience

• Server-Sent Events (SSE) for streaming AI responses
• Topic tracking system to ensure complete data collection
• Confidence scoring for extracted fields
• Graceful fallback to traditional forms when needed

Challenge 3: Vector Search at Scale

• Two-tier deployment model:
  • Cloud (netpad.io): M10+ Atlas cluster for Vector Search
  • Self-hosted: Atlas Local (Docker) for development and private instances
• Intelligent document chunking with sentence-aware boundaries
• Configurable retrieval parameters (chunk count, minimum score)

Challenge 4: Workflow Reliability

• MongoDB-based job queue with priority support
• Configurable retry policies with exponential backoff
• Per-node timeout management
• Execution history with full audit trail

Developer Ecosystem

NPM Packages

bash code-highlight# React form renderer (28+ field types)
npm install @netpad/forms

# Workflow API client
npm install @netpad/workflows

# CLI for package management
npm install -g @netpad/cli

# MCP server for AI-assisted development (75 tools)
npm install @netpad/mcp-server

165+ API Endpoints

• Form CRUD, submissions, and analytics
• Workflow management and execution
• Organization and project management
• Application marketplace and npm integration
• AI and RAG operations

Open Source & Deployment

Subscription Tiers

Real-World Applications

1. Workshop Registration: Collect attendee information with automatic confirmation workflows
2. Support Ticket Systems: Conversational intake that extracts structured ticket data
3. Healthcare Intake: HIPAA-compliant patient forms with field-level encryption
4. Vendor Assessments: RAG-powered forms that reference policy documents
5. Lead Qualification: AI conversations that qualify leads and route to appropriate teams
6. Event Management: Multi-page registration with payment integration workflows

Future Enhancements

1. Enhanced AI Agents: More specialized agents for compliance, translation, and insights
2. Workflow Templates: Expanded library of pre-built automation patterns
3. Advanced Analytics: Deeper insights into form performance and user behavior
4. Multi-Language Support: Internationalization for global deployments
5. Plugin System: Third-party extensions and integrations
6. Mobile SDK: Native mobile form rendering

Conclusion