Chapter 7: Backend Deployment Options (Free Tier Ready)

Chapter 7 is all about getting your AI backend online—so anyone in the world can use your model or app. We'll walk through free-tier deployment options like Hugging Face Spaces, Railway, and Render, and help you choose the one that fits your project best.

---

7.1 Before You Deploy — Checklist

You should already have:

• /backend/app/main.py (your FastAPI logic or app.py for Gradio)

• requirements.txt with all dependencies

• .env (for local secrets)

• Dockerfile or a deployment config (optional)

We’ll now deploy this backend to one of these platforms:

---

7.2 Option A: Hugging Face Spaces (Gradio or FastAPI)

✅ Pros:

• Super beginner-friendly

• Great for demo apps or model showcases

• Free GPU (on PRO) or CPU (on free tier)

🚫 Cons:

• Limited memory (2–6 GB)

• CPU only unless upgraded

• Best for Gradio or lightweight FastAPI

FastAPI Setup on HF Spaces app.py

   from fastapi import FastAPI
   app = FastAPI()
   @app.get("/")
   def root():
       return {"message": "Hello Hugging Face!"}

requirements.txt

   fastapi
   uvicorn
   python-dotenv
   openai

README.md

   ---
   title: My AI API
   emoji: 🤖
   colorFrom: gray
   colorTo: indigo
   sdk: docker
   ---
   # My AI App
   An API powered by FastAPI and OpenAI!

Dockerfile (if needed)

   FROM python:3.10
   WORKDIR /app
   COPY . /app
   RUN pip install -r requirements.txt
   CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860"]

Push to Hugging Face:

   git init
   git remote add origin https://huggingface.co/spaces/your-username/your-space
   git add .
   git commit -m "initial commit"
   git push -u origin main

---

7.3 Option B: Railway (FastAPI + OpenAI + Replicate)

✅ Pros:

• Perfect for FastAPI-based backends

• Easy GitHub integration

• 500 free compute hours/month

🚫 Cons:

• Cold starts (10–30s delay)

• No GPU on free tier

• Can timeout on long API responses

FastAPI Setup for Railway
backend/requirements.txt

   fastapi
   uvicorn
   python-dotenv
   openai
   replicate

pyproject.toml (optional if you need Railway to detect the start command)

   [tool.poetry]
   name = "myaiapp"
   version = "0.1.0"
   [tool.poetry.scripts]
   start = "uvicorn app.main:app --host 0.0.0.0 --port $PORT"

1. Push Backend Repo to GitHub

   git init
   git add .
   git commit -m "ready for deployment"
   git remote add origin https://github.com/<your-username>/<your-repo>
   git push -u origin main
   

2. Connect to Railway
   • Go to https://railway.app
   • New Project → Deploy from GitHub
   • Set environment variables:
       ○ OPENAI_API_KEY
       ○ REPLICATE_API_TOKEN
   • Done ✅

---

7.4 Option C: Render

✅ Pros:

• Simple, fast deploys

• Free up to 750 hrs/month

• Can run background tasks

🚫 Cons:

• Cold starts like Railway

• Slower startup than Railway

Basic Deploy Steps:

1. Create backend repo → push to GitHub

2. Go to https://render.com

3. New → Web Service → Connect your GitHub repo

4. Use uvicorn app.main:app --host 0.0.0.0 --port 10000 as Start Command

5. Set environment variables

6. Done ✅

---

7.5 Managing Environment Variables

Platform	Where to Add Them
Hugging Face	Settings > Secrets
Railway	Project > Variables
Render	Environment > Add Environment Variables

Add:

   OPENAI_API_KEY=sk-xxxx
   REPLICATE_API_TOKEN=r8_xxxx

Don’t commit .env to GitHub — keep it local or use .gitignore.

---

7.6 Deployment Checklist

• Backend runs uvicorn app.main:app

• requirements.txt is complete

• GitHub repo is pushed

• Environment variables added

• Deployment platform is selected (HF, Railway, or Render)

• Test /generate or /predict endpoint with Postman or browser

Chapter Summary

• You deployed your backend to the cloud!
• Hugging Face Spaces = great for demos
• Railway = great for FastAPI-powered APIs
• Render = great alternative with generous limits
• Your AI app is now globally accessible 🌐

---