Setup Guide

Quick Start

1. Install Dependencies

   npm install
   

2. Configure Environment Variables

   Create a .env.local file in the root directory:

   # Required: Choose one AI provider
   ANTHROPIC_API_KEY=your_anthropic_api_key_here
   # OR
   OPENAI_API_KEY=your_openai_api_key_here
   
   # Required: Choose one vector database
   PINECONE_API_KEY=your_pinecone_api_key
   PINECONE_INDEX=your_index_name
   PINECONE_ENVIRONMENT=your_environment
   
   # OR Supabase
   SUPABASE_URL=your_supabase_url
   SUPABASE_KEY=your_supabase_key
   
   # Optional: Analytics
   SUPABASE_DB_URL=your_database_url
   
   # App Config
   NEXT_PUBLIC_SITE_URL=http://localhost:3000
   

3. Run Development Server

   npm run dev
   

4. Open Browser Navigate to http://localhost:3000

Getting API Keys

Anthropic Claude API

1. Go to console.anthropic.com
2. Sign up or log in
3. Navigate to API Keys
4. Create a new API key
5. Copy and paste into .env.local

Pinecone (Vector Database)

1. Go to app.pinecone.io
2. Sign up for a free account
3. Create a new index:
   • Name: fountain-qa (or your choice)
   • Dimensions: 1536 (for OpenAI embeddings) or 1024 (for Claude)
   • Metric: cosine
4. Copy API key and index name to .env.local

Alternative: Supabase Vector

1. Go to supabase.com
2. Create a new project
3. Enable pgvector extension in SQL editor:

   CREATE EXTENSION IF NOT EXISTS vector;
   

4. Copy project URL and anon key to .env.local

Document Processing (Next Steps)

Once you have your source document:

1. Place Document

   • Put your Fountain Workflows document in data/source-document.docx

2. Process Document (when implemented)

   npm run process-doc
   

3. Generate Embeddings (when implemented)

   npm run generate-embeddings
   

Testing

Test the system with sample questions:

npm run test-queries

Troubleshooting

"ANTHROPIC_API_KEY is not set"

• Make sure .env.local exists in the root directory
• Verify the API key is correct
• Restart the dev server after adding environment variables

Vector Search Not Working

• The current implementation uses mock data
• You'll need to implement the actual vector search in lib/vector/search.ts
• Or use the simpler approach: put the entire document in Claude's context (200K tokens)

Build Errors

• Run npm install again
• Delete .next folder and rebuild
• Check Node.js version (requires 18+)

Current Status

✅ Completed:

• Project structure
• UI components
• API routes
• AI integration (Claude)
• Intent classification
• Basic query processing

⏳ To Do:

• Document processing pipeline
• Vector database integration
• Embedding generation
• Analytics/logging
• Response caching

Next Steps

1. For MVP (Simpler Approach):

   • Skip vector database initially
   • Put entire lead-relevant document in Claude's context
   • Use Claude 3.5 Sonnet with 200K token context
   • Faster to implement and iterate

2. For Production (Full RAG):

   • Implement document chunking
   • Set up vector database
   • Generate embeddings
   • Implement similarity search
   • Add caching layer

Support

If you encounter issues:

1. Check the console for error messages
2. Verify all environment variables are set
3. Ensure API keys have proper permissions
4. Review the README.md for more details