MCP Chat Support System

# ☁️ Cloudflare Deployment Status ## ✅ Frontend (Cloudflare Pages) **Status:** ✅ Deployed **URL:** https://main.clientsphere.pages.dev **Deployment URL:** https://abaa49a3.clientsphere.pages.dev ### Deployment Command: ```bash npm run build npx wrangler pages deploy dist --project-name=clientsphere ``` ### Environment Variables (Set in Cloudflare Pages Dashboard): ``` VITE_API_URL=https://mcp-backend.officialchiragp1605.workers.dev VITE_RAG_API_URL=https://your-rag-backend.onrender.com VITE_GOOGLE_CLIENT_ID=your-google-client-id ``` --- ## ✅ Cloudflare Workers (MCP Backend) **Status:** ✅ Deployed **URL:** https://mcp-backend.officialchiragp1605.workers.dev ### Deployment Command: ```bash cd mcp-backend npx wrangler deploy ``` --- ## ⚠️ RAG Backend - Cloudflare Limitation **Status:** ❌ Cannot deploy to Cloudflare Workers ### Why? The RAG backend is a Python FastAPI application with heavy ML dependencies: - `sentence-transformers` (embedding models) - `chromadb` (vector database) - `pymupdf`, `python-docx` (document processing) - Large model files and dependencies **Cloudflare Workers limitations:** - Python support is limited/beta - 10MB bundle size limit (our dependencies are much larger) - No persistent file system for vector database - CPU/memory limits too restrictive for ML workloads ### ✅ Recommended Deployment: **Deploy to Render or Railway** (see `DEPLOYMENT_GUIDE.md`) The RAG backend needs: - Full Python 3.11+ environment - Persistent storage for vector database - Sufficient CPU/memory for ML operations - Ability to install large Python packages ### Alternative: Cloudflare Workers Proxy If you want to keep everything on Cloudflare, you could: 1. Deploy RAG backend to Render/Railway 2. Create a Cloudflare Worker that proxies requests to the RAG backend 3. This gives you Cloudflare's CDN benefits while hosting the heavy backend elsewhere --- ## 🔄 Updating Deployments ### Update Frontend: ```bash npm run build npx wrangler pages deploy dist --project-name=clientsphere ``` ### Update Workers Backend: ```bash cd mcp-backend npx wrangler deploy ``` ### Update RAG Backend (on Render/Railway): - Push to GitHub (auto-deploys if configured) - Or manually redeploy from dashboard --- ## 📝 Notes - Frontend and Workers backend are fully deployed on Cloudflare - RAG backend must be deployed to Render/Railway due to technical limitations - All services can work together via their respective URLs - Update environment variables to connect all services