FastAPI with Open AI deployed on Render

Use this repo as a template to deploy a Python FastAPI service on Render that includes OpenAI endpoints.

Why use this template

📌 Includes two endpoints: /prompt and /prompt/stream

📡 /prompt/stream streams responses from OpenAI → FastAPI → Client

🚀 Deploy with one click

✅ Includes Pytest - Modify tests in test_main.py and run tests with the command pytest

Run Project Locally

Clone or fork repository

git clone https://github.com/waseemhnyc/fastapi-openai-render.git

Create a virutalenv and source the environment

python3 -m venv venv
source venv/bin/activate

Install the necessary libraries

pip install -r requirements.txt

Create a .env file and input your OpenAI API Key in the file

cp .env.example .env

Run local server

uvicorn main:app --host 0.0.0.0

Run Test

pytest

Manual Steps To Deploy

1. You may use this repository directly or create your own repository from this template if you'd like to customize the code.

2. Create a new Web Service on Render.

3. Specify the URL to your new repository or this repository.

4. Render will automatically detect that you are deploying a Python service and use pip to download the dependencies.

5. Specify the following as the Start Command.

   uvicorn main:app --host 0.0.0.0 --port $PORT

6. Add OPENAI_API_KEY as an environment variable.

7. Click Create Web Service.

Or Click to Deploy:

Frequently Asked Questions (FAQ)

Question: How can I customize the Click to Deploy button to point to my forked repository?

By updating the URL to your forked repository in this README file.

Question: How can I change the name of my web service?

You can change the name of your web service by modifying the `name` key in the `render.yaml` file.

  - type: web
    name: your-custom-service-name
    runtime: python