🚀 Building AI Agents with FastAPI + OpenAI

#fastapi #openai #python #cicd

AI agents are becoming an essential part of modern applications — whether it’s customer support, workflow automation, or intelligent assistants that reason and act.

In this post, we’ll build a simple AI agent API using FastAPI and OpenAI, showing how to structure it for real-world use.

🧠 What We’re Building

We’ll create a FastAPI app that:

Accepts a user query (e.g., “Summarize this article” or “Get today’s top tech news”).
Sends it to the OpenAI API.
Returns a structured AI-generated response.

You can later expand this into a multi-agent system or connect it with n8n, Zapier, or LangChain.

⚙️ Setup

Requirements:

pip install fastapi uvicorn openai python-dotenv

Create a .env file with your OpenAI key:

OPENAI_API_KEY=your_openai_api_key_here

🧩 The Code

main.py

from fastapi import FastAPI, HTTPException
from pydantic import BaseModel
from openai import OpenAI
import os
from dotenv import load_dotenv

load_dotenv()
app = FastAPI()
client = OpenAI(api_key=os.getenv("OPENAI_API_KEY"))

class Query(BaseModel):
    message: str

@app.post("/agent")
def run_agent(query: Query):
    try:
        completion = client.chat.completions.create(
            model="gpt-4o-mini",
            messages=[
                {"role": "system", "content": "You are a helpful AI agent."},
                {"role": "user", "content": query.message}
            ]
        )
        return {"response": completion.choices[0].message.content}
    except Exception as e:
        raise HTTPException(status_code=500, detail=str(e))

Run it with:

uvicorn main:app --reload

Now, send a POST request:

curl -X POST http://127.0.0.1:8000/agent \
    -H "Content-Type: application/json" \
    -d '{"message": "Write a haiku about FastAPI"}'

⚡ Example Response

{
  "response": "FastAPI flies,\nCode flows with lightning speed bright,\nPython dreams take flight."
}

💡 Takeaway

With FastAPI + OpenAI, you can spin up a production-ready AI agent in minutes — and scale it into something much smarter.

If you’re building something similar (like custom AI workflows or real-time assistants), drop a comment or connect — I’d love to hear your approach! 👇

DEV Community