Customer Registration MCP Server

# Quick Vercel Deployment ## ⚠️ Important Notice **Vercel deployment has limitations:** - ❌ MCP server cannot run (requires long-running process) - ❌ No actual customer creation via API - ❌ No conversation history - ✅ Beautiful UI works perfectly - ✅ OpenAI conversations work **For full functionality, use Railway, Render, or VPS deployment.** See [DEPLOYMENT.md](./DEPLOYMENT.md) --- ## Deploy to Vercel (Demo Mode) ### Step 1: Push to GitHub ```bash git push origin main ``` ### Step 2: Import to Vercel 1. Go to [vercel.com](https://vercel.com) 2. Click **"Add New Project"** 3. Import your GitHub repository: `rodrigoai/mcpNova` 4. Vercel will auto-detect the configuration ### Step 3: Configure Environment Variables Add these in Vercel dashboard → Settings → Environment Variables: **Required:** ``` OPENAI_API_KEY=sk-proj-... ``` **Optional:** ``` AGENT_TONE=Professional, helpful, and efficient ``` ### Step 4: Deploy - Click **"Deploy"** - Wait for build to complete (~2-3 minutes) - Visit your deployment URL: `https://mcp-nova-xxx.vercel.app` --- ## What Works on Vercel ✅ **Beautiful responsive UI** ✅ **Chat interface** ✅ **OpenAI conversations** ✅ **Data collection** ✅ **Auto-deploy on push** ❌ **Customer creation** (MCP not available) ❌ **Conversation persistence** --- ## Testing Your Deployment Open your Vercel URL and try: ``` User: Hello! Bot: [Responds with greeting] User: I want to register a customer Bot: [Collects information but cannot create customer] ``` The bot will collect data but display a message about serverless limitations. --- ## Viewing Logs ```bash # Install Vercel CLI npm i -g vercel # Login vercel login # View logs vercel logs ``` --- ## For Full Functionality Deploy to platforms that support long-running processes: ### Option 1: Railway (Recommended - Easiest) 1. Go to [railway.app](https://railway.app) 2. Connect GitHub repo 3. Add environment variables (including CUSTOMER_API_*) 4. Deploy automatically 5. Get full MCP integration ✅ ### Option 2: Render.com 1. Go to [render.com](https://render.com) 2. New Web Service → Connect repo 3. Build: `yarn build` 4. Start: `yarn chatbot:start` 5. Add all environment variables 6. Deploy ✅ See [DEPLOYMENT.md](./DEPLOYMENT.md) for detailed instructions. --- ## Environment Variables for Full Deployment ```env # Required OPENAI_API_KEY=sk-... CUSTOMER_API_HOST=https://your-api.com CUSTOMER_API_TOKEN=your-bearer-token # Optional AGENT_TONE=Professional, helpful, and efficient CHATBOT_PORT=3000 NODE_ENV=production ``` --- ## Cost Comparison | Platform | Free Tier | Full MCP | Auto-Scale | Ease | |----------|-----------|----------|------------|------| | Vercel | ✅ Yes | ❌ No | ✅ Yes | ⭐⭐⭐⭐⭐ | | Railway | ✅ Limited | ✅ Yes | ✅ Yes | ⭐⭐⭐⭐⭐ | | Render | ✅ Yes | ✅ Yes | ✅ Yes | ⭐⭐⭐⭐ | | AWS EC2 | ❌ No | ✅ Yes | ⚠️ Manual | ⭐⭐ | --- ## Troubleshooting ### Build fails on Vercel - Check `package.json` has correct dependencies - Ensure TypeScript compiles: `yarn build` - Check Vercel build logs ### Chat doesn't work - Verify `OPENAI_API_KEY` is set in Vercel - Check browser console for errors - Test API: `curl https://your-url.vercel.app/api/chat` ### Want customer creation to work? - Deploy to Railway/Render instead - See [DEPLOYMENT.md](./DEPLOYMENT.md) --- ## Next Steps 1. ✅ Deploy to Vercel for UI demo 2. ✅ Test the interface 3. 📝 Deploy to Railway for full functionality 4. 🎉 Use in production --- ## Support - Full deployment guide: [DEPLOYMENT.md](./DEPLOYMENT.md) - UI documentation: [UI_GUIDE.md](./UI_GUIDE.md) - Quick start: [CHATBOT_QUICKSTART.md](./CHATBOT_QUICKSTART.md)