DEPLOY.md

Talking Avatar with AI - Deployment Guide

🚀 Deployment Options

This project is configured to be deployed as a single Docker container, which hosts both the Node.js backend and the React frontend.

Option 1: Render (Recommended)

Render is a unified cloud to build and run all your apps and websites.

1. Create Render Account

   • Go to https://render.com
   • Sign up with GitHub

2. Create New Web Service

   • Click "New +" -> "Web Service"
   • Connect your GitHub repository
   • Render should automatically detect the render.yaml or Dockerfile.
   • If asked, choose Docker as the runtime.

3. Configuration

   • Name: healbot-avatar (or your choice)
   • Region: Choose one close to you
   • Branch: main
   • Runtime: Docker
   • Plan: Free (or higher)

4. Deploy

   • Click "Create Web Service"
   • Render will build the Docker image (this may take a few minutes) and deploy it.

Option 2: Hugging Face Spaces

You can deploy this application to Hugging Face Spaces using their Docker SDK.

1. Create Hugging Face Account

   • Go to https://huggingface.co
   • Sign up

2. Create New Space

   • Click on your profile -> "New Space"
   • Space Name: healbot-avatar
   • License: MIT (optional)
   • SDK: Docker
   • Hardware: CPU Basic (Free) is usually sufficient, but Upgrade if needed for faster TTS.

3. Upload Code

   • You can sync your GitHub repo with the Space, or push your code directly to the Space's repo.
   • Ensure the Dockerfile is in the root directory.

4. Configuration

   • Hugging Face Spaces usually listen on port 7860. You might need to update the Dockerfile or server.js if HF doesn't map port 3000 automatically.
   • Note: For Hugging Face, it is recommended to change the port in apps/backend/server.js to 7860 or use an environment variable.
   • Add a variable PORT = 7860 in the Space Settings -> "Variables and secrets".

📋 Pre-Deployment Checklist

• ✅ Removed Vercel and Railway configurations
• ✅ Added Dockerfile for multi-stage build
• ✅ Added render.yaml for Render Blueprints
• ✅ Backend serves Frontend static files
• ✅ edge-tts and ffmpeg included in Docker image

🔧 Environment Variables

Render / Hugging Face:

• PORT: Defaults to 3000. Set to 7860 for Hugging Face if needed.

🧪 Testing After Deployment

1. Open your deployed URL (e.g., https://healbot-avatar.onrender.com or https://huggingface.co/spaces/username/healbot-avatar)
2. The frontend should load.
3. Send a message. The backend should process it using edge-tts and return the audio and animation.

🆘 Troubleshooting

• Build Fails: Check the logs. Ensure ffmpeg and python are correctly installed in the Docker image.
• Audio Issues: If TTS fails, check if edge-tts is working. The logs will show python errors.
• Connection Refused: Ensure the port matches the environment variable.