Skip to content

DEPLOY_NOW.md

Latest commit

 

History

History
162 lines (114 loc) · 4.12 KB

DEPLOY_NOW.md

File metadata and controls

162 lines (114 loc) · 4.12 KB

🚀 Deploy Gemini Emotion Analysis - Step by Step

✅ Good News!

Your Gemini API key is working perfectly! The test showed it can detect:

  • joy - for happy/excited responses
  • curiosity - for questions and exploration
  • calm - for empathetic/neutral responses
  • And more emotions based on context!

🎯 What You Need to Do

You have 2 options to deploy. Choose the one that's easiest for you:

Option A: Supabase Dashboard (Recommended - No CLI needed)

Step 1: Set the API Key Secret

  1. Go to https://app.supabase.com
  2. Select your project
  3. Click Project Settings (gear icon in sidebar)
  4. Go to Edge Functions section
  5. Click Manage secrets or Add secret
  6. Add new secret:
    • Name: GEMINI_API_KEY
    • Value: AIzaSyALKf9SnATD_4OEh_Atx7AFXDSj6hhpfho
  7. Click Save or Add

Step 2: Deploy the Function

  1. In Supabase Dashboard, go to Edge Functions
  2. Look for gemini-chat function (or click Create function)
  3. Click Edit or Create new function
  4. Name it: gemini-chat
  5. Copy ALL the code from this file: 
    supabase/functions/gemini-chat/index.ts
  6. Paste into the function editor
  7. Click Deploy or Save

Step 3: Test It!

  1. Refresh your web app (http://localhost:5173)
  2. Try these messages:
    • "This is amazing!" → Should show joy 🟡
    • "How does this work?" → Should show curiosity 🔵
    • "I understand" → Should show calm 🔵
    • "This is frustrating" → Should show anger 🔴

Option B: Using Supabase CLI

Step 1: Install Supabase CLI

npm install -g supabase

Step 2: Login and Link

# Login to Supabase
supabase login

# Link to your project (get project-ref from dashboard URL)
supabase link --project-ref YOUR_PROJECT_REF

Step 3: Deploy Using Script

# Run the deployment script
.\deploy-function.ps1

Or manually:

# Set the API key
supabase secrets set GEMINI_API_KEY=AIzaSyALKf9SnATD_4OEh_Atx7AFXDSj6hhpfho

# Deploy the function
supabase functions deploy gemini-chat

🔍 Verify It's Working

After deployment, check:

  1. Browser Console (F12) - Should see:

    Gemini emotion analysis: { emotion: 'joy', confidence: 0.9, ... }

  2. Supabase Function Logs - Go to Edge Functions → gemini-chat → Logs

  3. Visual Changes - The AI avatar should change colors based on different emotions

📊 What Changed

Before (Keyword-based):

"How does this work?" → Always "curiosity" (keyword match)
"This is great!" → Always "curiosity" (default)
"I understand" → Always "curiosity" (default)

After (Gemini-powered):

"How does this work?" → "curiosity" (asking question)
"This is great!" → "joy" (positive excitement)
"I understand" → "calm" (acknowledgment)
"I'm confused" → "confusion" (uncertainty)

⚠️ Troubleshooting

Still showing same emotion?

  1. Clear browser cache: Ctrl+Shift+Delete
  2. Hard reload: Ctrl+F5
  3. Check Supabase logs: Look for errors in function logs
  4. Verify API key: Make sure it's set in Supabase secrets (not local .env)

API Errors?

  1. Check Gemini API quota: https://makersuite.google.com/app/apikey
  2. Verify API key is correct in Supabase secrets
  3. Check function logs for specific error messages

Function not deploying?

  1. Make sure you copied the ENTIRE file content
  2. Check for syntax errors in the editor
  3. Try deploying via CLI instead of dashboard

📝 Files Updated

✅ supabase/functions/gemini-chat/index.ts - Added Gemini emotion analysis
✅ src/hooks/useAI.ts - Uses backend emotion instead of local analysis
✅ Test script confirms API is working

🎉 Next Steps

  1. Deploy using one of the options above
  2. Test with different message types
  3. Watch the AI's emotions change dynamically!

Need help? Check the Supabase function logs for detailed error messages.