Skip to content

DEPLOYMENT_GUIDE.md

Latest commit

 

History

History
94 lines (77 loc) · 4.11 KB

DEPLOYMENT_GUIDE.md

File metadata and controls

94 lines (77 loc) · 4.11 KB
Raw

🚀 Deployment Guide: AI Interviewer

This guide explains how to set up the AI Interviewer locally and how to deploy it to the cloud for free.

🛠️ Prerequisites

  1. Node.js & npm installed.
  2. Python 3.10+ installed.
  3. A GitHub account.
  4. An Atlas MongoDB account (Free tier).
  5. An Upstash Redis account (Free tier, required for background queues).
  6. A Google AI Studio account (for Gemini API Key).
  7. A Groq Console account (for Groq API Key for transcription).
  8. A Google Cloud Console project (for Google Login/OAuth).
  9. A JDoodle account (for Live Code Execution API).

💻 Local Development Setup

1. Backend (/backend)

  1. Create a .env file: 
    MONGO_URI=your_mongodb_connection_string_here
FRONTEND_URL=http://localhost:5173
PORT=5000
NODE_ENV=development
JWT_SECRET=your_secure_jwt_secret_here
GOOGLE_CLIENT_ID=your_google_client_id_here
GOOGLE_CLIENT_SECRET=your_google_client_secret_here
AI_SERVICE_URL=http://localhost:8000
JDOODLE_CLIENT_ID=your_jdoodle_client_id_here
JDOODLE_CLIENT_SECRET=your_jdoodle_client_secret_here
REDIS_URL=rediss://default:your_upstash_redis_password@your-upstash-endpoint.upstash.io:6379
CLOUDINARY_CLOUD_NAME=your_cloudinary_cloud_name_here
CLOUDINARY_API_KEY=your_cloudinary_api_key_here
CLOUDINARY_API_SECRET=your_cloudinary_api_secret_here
  2. Run: npm install && npm run dev

2. AI Service (/ai-service)

  1. Create a .env file: 
    PORT=8000
MODEL_NAME=gemini-3.1-flash-lite
GEMINI_API_KEY=your_gemini_api_key_here
GROQ_API_KEY=your_groq_api_key_here
ALLOWED_ORIGINS=http://localhost:5000,http://localhost:5173
REQUEST_TIMEOUT=60
  2. Run: pip install -r requirements.txt && python main.py

3. Frontend (/frontend)

  1. Create a .env file: 
    VITE_API_URL=http://localhost:5000/api
VITE_GOOGLE_CLIENT_ID=your_google_client_id_here
  2. Run: npm install && npm run dev

☁️ Production Deployment (Free Tier)

1. 🚀 Deploy Backend & AI Service (Render)

We have provided a render.yaml Blueprint which automatically configures Docker-based deployments for both services. The AI Service uses Docker to properly install tesseract-ocr and poppler-utils.

  1. Go to your Render Dashboard and click New -> Blueprint.
  2. Connect your GitHub repository.
  3. Render will detect the render.yaml file.
  4. Fill in the required environment variables in the Render UI (e.g., MONGO_URI, JWT_SECRET, GOOGLE_CLIENT_ID, GOOGLE_CLIENT_SECRET, FRONTEND_URL, GEMINI_API_KEY, GROQ_API_KEY, REDIS_URL).
  5. Click Apply. Render will build and deploy both the ai-interviewer-backend and ai-interviewer-ai-service automatically using the provided Dockerfiles.

Note: You can find the AI Service URL in your Render dashboard after deployment. Update your Backend's AI_SERVICE_URL environment variable to point to it, and the Backend URL will go in your Frontend's VITE_API_URL.

3. 💻 Frontend (Vercel)

  • Import Project -> Connect Repo.
  • Root Directory: frontend
  • Framework Preset: Vite
  • Env Vars: Add VITE_API_URL (your Render backend URL + /api) and VITE_GOOGLE_CLIENT_ID.

🔗 Finalizing Google OAuth

To make Google Login work in production:

  1. Go to Google Cloud Console -> APIs & Services -> Credentials.
  2. Add your Vercel URL (e.g., https://my-app.vercel.app) to Authorized JavaScript Origins.
  3. Add https://your-backend.onrender.com/api/user/auth/google/callback to Authorized redirect URIs.

💡 Troubleshooting

  • 502 Bad Gateway: Usually a "Cold Start" on Render. Wait 30 seconds and refresh.
  • Cookie/Login Issues: Ensure NODE_ENV=production is set in your Render Backend environment variables.
  • Memory Limits: This app is optimized to run on 512MB RAM using Groq's blazing fast Whisper API for audio. Do not use local Whisper/Torch models on Render Free Tier.