Skip to content

QUICK_START.md

Latest commit

 

History

History
148 lines (113 loc) · 3.72 KB

QUICK_START.md

File metadata and controls

148 lines (113 loc) · 3.72 KB

Quick Start Guide - AI Interview Mocker

✅ Current Status

Your project is ready to run with:

  • ✅ Ollama installed (v0.13.0)
  • ✅ Phi-3 Mini model available (2.2 GB)
  • ✅ Database configured
  • ✅ Authentication set up
  • ✅ AI configured to use local Ollama (FREE!)

🚀 Run the Project (3 Steps)

Step 1: Install Dependencies

npm install

Step 2: Set Up Database

npm run db:push

Step 3: Start Development Server

npm run dev

That's it! Open http://localhost:3000

📱 How to Use

  1. Sign Up/Login - Click "Start Mock Interview"
  2. Create Interview - Fill in:
    • Job Position (e.g., "Full Stack Developer")
    • Tech Stack (e.g., "React, Node.js, MongoDB")
    • Years of Experience (e.g., "3")
  3. AI Generates Questions - Ollama creates 5 personalized questions
  4. Take Interview - Answer using voice or text
  5. Get Feedback - Receive AI-powered ratings and suggestions

🔧 Troubleshooting

Ollama not working?

# Check if running
curl http://localhost:11434/api/tags

# Start Ollama
ollama serve

# Test AI
node utils/testAI.js

Database issues?

# Reset database
npm run db:push

Port 3000 already in use?

# Kill the process
npx kill-port 3000

# Or use different port
npm run dev -- -p 3001

🎯 For Your PPT Demo

Demo Flow:

  1. Show landing page (professional UI)
  2. Sign up/login (Clerk authentication)
  3. Create interview (AI generates questions in ~5 seconds)
  4. Show interview interface (questions + voice recording)
  5. Answer a question (speech-to-text in action)
  6. Show feedback page (AI analysis with ratings)

Key Points to Highlight:

  • Local AI - No API costs, unlimited usage
  • Privacy - All data stays on your machine
  • Real-time - Instant speech-to-text conversion
  • Intelligent - AI-powered feedback and ratings
  • Modern Stack - Next.js, React, Tailwind, PostgreSQL

📊 What Makes This Special

  1. Free AI - Uses local Ollama instead of paid APIs
  2. Voice Enabled - Speech-to-text for natural interaction
  3. Personalized - Questions tailored to your role and experience
  4. Instant Feedback - No waiting for human reviewers
  5. Progress Tracking - See your improvement over time

🎓 For Your Presentation

Slide Suggestions:

  1. Problem - Interview preparation is expensive and time-consuming
  2. Solution - AI-powered mock interviews with local processing
  3. Tech Stack - Modern, scalable, and cost-effective
  4. Demo - Live walkthrough of the application
  5. Impact - Free, unlimited practice for job seekers

Technical Highlights:

  • Next.js 14 with App Router
  • Local AI with Ollama (Phi-3 Mini)
  • Real-time speech recognition
  • PostgreSQL with Drizzle ORM
  • Clerk authentication
  • Responsive design with Tailwind CSS

💡 Pro Tips

For Best Demo:

  • Test your microphone before presenting
  • Have a sample job description ready
  • Show the feedback page (most impressive part)
  • Mention the cost savings (free vs paid APIs)

For Development:

  • Use npm run db:studio to view database
  • Check Ollama logs: ollama logs
  • Monitor performance: ollama ps

🔄 Alternative: Switch to Gemini

If you want to use Gemini instead:

  1. Get API key from https://makersuite.google.com/app/apikey
  2. Update .env.local: 
    NEXT_PUBLIC_AI_PROVIDER=gemini
NEXT_PUBLIC_GEMINI_API_KEY=your_key_here
  3. Restart server

But Ollama is recommended because it's free and works great!

📞 Need Help?

  • Check AI_SETUP.md for detailed AI configuration
  • Check GEMINI_VS_OLLAMA.md for comparison
  • Run node utils/testAI.js to test AI connection

You're all set! Your AI Interview Mocker is ready to impress. 🚀