Serving apps

Flyte SDK lets you serve apps on your Union/Flyte instance, making them accessible via HTTP endpoints. Apps are long-running services that can be accessed by users or other services.

First install FastAPI in your virtual environment:

pip install fastapi

Hello world example

Create a file called hello_app.py with the following content:

# /// script
# requires-python = "==3.13"
# dependencies = [
#    "flyte==2.0.0b31",
#    "fastapi",
#    "uvicorn",
# ]
# ///

"""A simple "Hello World" FastAPI app example for serving."""

from fastapi import FastAPI
import pathlib
import flyte
from flyte.app.extras import FastAPIAppEnvironment

# Define a simple FastAPI application
app = FastAPI(
    title="Hello World API",
    description="A simple FastAPI application",
    version="1.0.0",
)

# Create an AppEnvironment for the FastAPI app
env = FastAPIAppEnvironment(
    name="hello-app",
    app=app,
    image=flyte.Image.from_debian_base(python_version=(3, 12)).with_pip_packages(
        "fastapi",
        "uvicorn",
    ),
    resources=flyte.Resources(cpu=1, memory="512Mi"),
    requires_auth=False,
)

# Define API endpoints
@app.get("/")
async def root():
    return {"message": "Hello, World!"}

@app.get("/health")
async def health_check():
    return {"status": "healthy"}

# Serving this script will deploy and serve the app on your Union/Flyte instance.
if __name__ == "__main__":
    # Initialize Flyte from a config file.
    flyte.init_from_config(root_dir=pathlib.Path(__file__).parent)

    # Serve the app remotely.
    app_instance = flyte.serve(env)

    # Print the app URL.
    print(app_instance.url)
    print("App 'hello-app' is now serving.")

Understanding the code

In the code above we do the following:

• Import the flyte package and FastAPIAppEnvironment from flyte.app.extras.
• Define a FastAPI application using the FastAPI class.
• Create an AppEnvironment using FastAPIAppEnvironment:
  • Apps are long-running services, unlike tasks which run to completion.
  • The FastAPIAppEnvironment automatically configures the app to run with uvicorn.
  • We specify the container image with required dependencies (FastAPI and uvicorn).
  • We set resource limits (CPU and memory).
  • We disable authentication for this example (requires_auth=False) so you can easily access the app with a curl command.

Serving the app

Make sure that your config.yaml file is in the same directory as your hello_app.py script.

Now, serve the app with:

flyte serve hello_app.py env

You can also serve it via python:

python hello_app.py

This will use the code in the if __name__ == "__main__": block to serve the app with the flyte.serve() function.

You can also serve the app using python hello_app.py, which uses the main guard section in the script. It invokes flyte.init_from_config() to set up the connection with your Union/Flyte instance and flyte.serve() to deploy and serve your app on that instance.

Viewing the results

In your terminal, you should see output like this:

https://my-instance.flyte.com/v2/domain/development/project/flytesnacks/apps/hello-app
App 'hello-app' is now serving.

Click the link to go to your Union instance and see the app in the UI, where you can find the app URL, or visit /docs for the interactive Swagger UI API documentation.

Next steps

Now that you’ve served your first app, you can learn more about:

• Configuring apps: Learn how to configure app environments, including images, resources, ports, and more
• Building apps: Explore different types of apps you can build (FastAPI, Streamlit, vLLM, SGLang)
• Serving and deploying apps: Understand the difference between serving (development) and deploying (production) apps