DocuMind - AI-Powered Knowledge Management System

Build an intelligent Q&A system for company documents using Claude, RAG, and agentic engineering.

What is DocuMind?

DocuMind is an AI chatbot that answers questions from your company's documents. Ask questions in natural language, get accurate answers with source citations.

What You'll Build (Sessions 3-10)

Session	Topic	What You'll Create
S3	Skills, Subagents, Hooks	Document upload pipeline
S4	Database & MCP	Supabase integration
S5	Multi-Agent Systems	Parallel document processing
S6	RAG Implementation	Q&A interface
S7	Advanced Parsing	PDF/DOCX support
S8	Vector Databases	Semantic search
S9	Memory & Learning	Conversation memory
S10	Evaluation	Quality metrics

Quick Start

1. Fork this repository to your GitHub account
2. Launch in Codespaces: Click green Code button → Codespaces → Create codespace on main
3. Follow setup guide: See docs/guides/SETUP.md
4. Get workshop files from your instructor

Prerequisites

• Basic Python & JavaScript knowledge
• GitHub account (free)
• Anthropic API key (free tier available)
• OpenAI API key (pay-as-you-go)

Project Structure

heroforge-documind/
├── .claude/              # Claude Code configuration
│   ├── agents/          # AI agent definitions
│   ├── skills/          # Custom skills (you'll create these!)
│   ├── subagents/       # Specialized agents (you'll create these!)
│   └── hooks/           # Automation scripts (you'll create these!)
├── src/
│   └── documind/        # Main DocuMind application code
├── tests/               # Test files
├── docs/
│   └── guides/          # Setup and troubleshooting guides
├── .env.example         # Template for environment variables
└── package.json         # Node.js dependencies

Technologies

• Claude Code - AI-powered development assistant
• Supabase - PostgreSQL database with pgvector for semantic search
• OpenRouter - Multi-model LLM access (GPT-4, Gemini, etc.)
• RAGAS - Evaluation framework for RAG systems

Environment Setup

# 1. Copy environment template
cp .env.example .env

# 2. Add your API keys to .env file

# 3. Install dependencies
npm install

# 4. Run tests
npm test

# 5. Launch Claude Code
dsp

Getting Help

• Workshop issues: Raise your hand or ask your instructor
• Technical bugs: Create a GitHub Issue in your fork
• Course questions: Ask in the course chat/Discord

License

MIT License - See LICENSE file