MCP Quoting System

# 🚀 Now What? - Getting Started Guide ## You've Built It - Now Let's Use It! Your MCP Quoting System is ready to help you **automatically compare new RFQs against your historical quotes** to generate accurate cost estimates and lead times. --- ## 🎯 What This System Does **The Problem It Solves:** When a customer sends you an RFQ (Request for Quote), you need to: 1. Figure out if you've made something similar before 2. Estimate the cost based on past work 3. Calculate lead time 4. Generate a professional quote **What This System Does:** 1. ✅ Reads the RFQ text 2. ✅ Searches your historical quotes database 3. ✅ Finds similar past work (with similarity scores) 4. ✅ Estimates cost using activity-based costing 5. ✅ Calculates lead time 6. ✅ Generates a quote document 7. ✅ Shows confidence level (high/medium/low) **Time Saved:** What used to take 30-60 minutes now takes 2-3 minutes! --- ## 🏃 Quick Start - 3 Simple Steps ### Step 1: Start the Server (10 seconds) Double-click **`START.bat`** and choose **[2] Start Development Server** The server will start on **http://localhost:3789** You'll see: ``` MCP QUOTING SYSTEM - Development Mode Server running on http://localhost:3789 Ready to receive requests! ``` --- ### Step 2: Send Your First RFQ (30 seconds) **Option A: Use the Quick Test (Easiest)** 1. Keep START.bat menu open 2. Press **[4] Quick Test** 3. Watch it automatically test with sample RFQs 4. See the results with match scores and estimates! **Option B: Use the Browser** 1. Open your browser to: http://localhost:3789/mcp/utility/historicalQuotes 2. You'll see your 5 sample historical quotes **Option C: Test with Your Own RFQ (Most Useful)** 1. Open `test-rfq.json` in Notepad 2. Edit the RFQ text with your own request 3. Save the file 4. Run `quick-test.bat` 5. See the results! --- ### Step 3: Review the Results You'll get back: - **Parsed RFP:** What the system understood from your RFQ - **Similar Quotes Found:** Past quotes ranked by similarity - **Match Score:** How similar (85%+ = high confidence) - **Cost Estimate:** Price per unit and total - **Lead Time:** Estimated days to deliver - **Confidence Level:** High/Medium/Low - **Quote Document:** Draft quote ready for review --- ## 📋 Real-World Usage Scenarios ### Scenario 1: Customer Emails You an RFQ **What You Receive:** ``` "Hi, we need 250 pieces of 6061-T6 aluminum brackets, CNC machined with anodized finish. Tolerance +/-0.005. Need them by March 15th." ``` **What You Do:** 1. **Copy the text** from the email 2. **Open PowerShell or Git Bash** and run: ```bash curl -X POST http://localhost:3789/mcp/invoke/evaluateRfpAndDraftQuote \ -H "Content-Type: application/json" \ -d "{\"rfp\":{\"rawText\":\"We need 250 pieces of 6061-T6 aluminum brackets, CNC machined with anodized finish. Tolerance +/-0.005. Need them by March 15th.\",\"qty\":250,\"contactEmail\":\"customer@example.com\"}}" ``` 3. **Review the response:** - Match found: Quote Q-1001 (92% similar!) - Estimated cost: $12.75/unit = $3,187.50 total - Lead time: 21 days - Confidence: HIGH 4. **Make your decision:** - High confidence = Quote is reliable, send to customer - Medium confidence = Review and adjust pricing - Low confidence = Needs engineer review --- ### Scenario 2: Planning a Quote for a Customer Meeting **Before the meeting:** 1. Start the server 2. Enter 3-4 variations of what the customer might ask for 3. Get instant estimates for each scenario 4. Go to meeting prepared with ballpark numbers **Example:** ```bash # Scenario 1: Small quantity curl -X POST http://localhost:3789/mcp/invoke/evaluateRfpAndDraftQuote \ -H "Content-Type: application/json" \ -d "{\"rfp\":{\"rawText\":\"50 aluminum parts, CNC\",\"qty\":50}}" # Scenario 2: Large quantity curl -X POST http://localhost:3789/mcp/invoke/evaluateRfpAndDraftQuote \ -H "Content-Type: application/json" \ -d "{\"rfp\":{\"rawText\":\"500 aluminum parts, CNC\",\"qty\":500}}" ``` --- ### Scenario 3: Reviewing Historical Data **Want to see what you've quoted before?** Open browser to: **http://localhost:3789/mcp/utility/historicalQuotes** You'll see all your past quotes with: - Customer names - Materials used - Processes required - Costs and lead times - Actual vs. estimated lead times **Use this to:** - Identify pricing trends - Find similar past work quickly - Check if you're being consistent - Learn from past quotes --- ## 🎨 Integration Options ### Option 1: Manual Use (What You Have Now) - Keep server running - Use curl/PowerShell to send RFQs - Review results - Generate quotes manually **Best for:** Small volume, manual quote process --- ### Option 2: Email Integration (Future Enhancement) **What it could do:** - Monitor email inbox - Auto-detect RFQ emails - Process them automatically - Send draft quotes for review **Not built yet, but possible!** --- ### Option 3: Web Interface (Future Enhancement) **What it could look like:** - Upload RFQ documents - See results in nice dashboard - Edit estimates before sending - Track quote status **Not built yet, but on the roadmap!** --- ### Option 4: API Integration with Your Tools **Connect to:** - Your CRM (Salesforce, HubSpot) - Your quoting tool - Your ERP system - Your email system **How:** Use the REST API endpoints already built! --- ## 📊 Understanding the Output ### What the System Returns ```json { "parsedRfp": { "id": "abc123", "material": "6061-t6", "processes": ["cnc", "anodize"], "qty": 200, "confidence": "high" }, "matches": [ { "quote": { "id": "Q-1001", ... }, "score": 0.92, "matchingFields": ["material", "processes", "quantity"], "notes": "Same part family, same finish" } ], "estimate": { "pricePerUnit": 12.50, "totalPrice": 2500.00, "leadDays": 21, "confidence": "high", "breakdown": { "materialCost": 400, "processCost": 8.50, "toolingPerUnit": 1.00, "overhead": 1.43, "marginPct": 0.20 } }, "doc": { "quoteId": "QOUT-1234567890", "to": "customer@example.com", "lines": [...], "terms": "Net 30. Lead time estimate: 21 days", "confidence": "high", "status": "draft" } } ``` ### Key Metrics to Watch **Match Score:** - **85-100%:** Very similar to past work - HIGH confidence - **70-85%:** Similar family - MEDIUM confidence - **Below 70%:** New type of work - LOW confidence, engineer review needed **Confidence Level:** - **HIGH:** Send quote to customer (maybe with small adjustment) - **MEDIUM:** Review pricing, adjust if needed - **LOW:** Engineer needs to review thoroughly --- ## 🔧 Common Tasks ### Task: Add a New Historical Quote After you complete a job, add it to the database: ```bash curl -X POST http://localhost:3789/mcp/utility/addHistoricalQuote \ -H "Content-Type: application/json" \ -d '{ "id": "Q-NEW", "quoteDate": "2024-11-13T10:00:00Z", "customerName": "Acme Corp", "normalized": { "material": "6061-t6", "processes": ["cnc", "anodize"], "qtyRange": [101, 500], "tolerances": "+/-0.005" }, "costPerUnit": 12.50, "totalCost": 2500.00, "leadDays": 21, "approved": true }' ``` --- ### Task: View All Your Quotes Browser: http://localhost:3789/mcp/utility/historicalQuotes Or command line: ```bash curl http://localhost:3789/mcp/utility/historicalQuotes ``` --- ### Task: Test Different Quantities ```bash # Small order curl -X POST http://localhost:3789/mcp/invoke/evaluateRfpAndDraftQuote \ -H "Content-Type: application/json" \ -d "{\"rfp\":{\"rawText\":\"25 pcs aluminum CNC\",\"qty\":25}}" # Large order curl -X POST http://localhost:3789/mcp/invoke/evaluateRfpAndDraftQuote \ -H "Content-Type: application/json" \ -d "{\"rfp\":{\"rawText\":\"1000 pcs aluminum CNC\",\"qty\":1000}}" ``` --- ## 📈 Next Steps - Making It Your Own ### 1. Add Your Historical Quotes - Edit `data/quotes.json` - Add your real past quotes - Include materials you actually work with - Add processes you actually do ### 2. Adjust Pricing - Edit `src/config.ts` - Update material costs - Update machine hour rates - Update overhead percentage - Update margin percentage ### 3. Test with Real RFQs - Take recent customer RFQs - Run them through the system - Compare estimates to your actual quotes - Tune the pricing until it matches ### 4. Build Your Workflow - Decide: Manual process or integration? - Train your team how to use it - Document your quote approval process - Set up regular database updates --- ## 🎓 Learn More - **TEST-CASES.md** - 7 example RFQs to try - **QUICKSTART.md** - Visual guide with examples - **README.md** - Complete API documentation - **BATCH-FILES-README.md** - All the shortcuts --- ## 💡 Pro Tips 1. **Keep the server running** during business hours 2. **Test with real RFQs** to build confidence 3. **Update historical quotes** after completing jobs 4. **Review medium/low confidence** estimates before sending 5. **Adjust pricing in config.ts** to match your actual costs 6. **Add more historical quotes** = better matching 7. **Use it for ballpark estimates** in customer calls --- ## 🆘 Need Help? ### Server won't start? - Check if port 3789 is available - Run `stop.bat` first - Check for error messages ### Results don't make sense? - Check your historical quotes - Review pricing in `src/config.ts` - Try with different quantities ### Want to customize? - Edit `src/config.ts` for pricing - Edit `data/quotes.json` for historical data - Edit `src/matcher.ts` for matching logic --- ## 🎯 Your Mission (If You Choose to Accept It) **Today:** 1. ✅ Start the server 2. ✅ Run the quick test 3. ✅ Try one real RFQ from your inbox **This Week:** 1. Add 10-20 real historical quotes 2. Test with 5-10 real customer RFQs 3. Tune pricing to match your actuals **This Month:** 1. Build it into your daily workflow 2. Train your team 3. Track time savings 4. Celebrate! 🎉 --- **Ready?** Double-click **START.bat** and press **[2]** to begin! 🚀 **Questions?** Just ask - I'm here to help!