Deploy a Studio

Turn your Studio into a production-grade server - no Docker, no YAML.

Manually packaging servers into containers slows down iteration and forces AI builders to learn DevOps. Snapshot Deploy eliminates that overhead by turning your working Studio into a real deployment - no Docker, no YAML, just one click to production.

Once your live server is working, you can instantly turn it into a production deployment. Lightning AI snapshots the full Studio environment (code, packages, ports, everything) and launches it as a scalable, monitored deployment.

Deploy via the UI

Initial Deployment

How it works:

• Install the Deployment plugin in the Studio
• Click “New Deployment”, select “Studio” as source
• Type the same command you'd use to run the server
• Click Deploy

Lightning AI handles:

• Packaging the environment
• Launching a deployment
• Monitoring traffic, autoscaling, logging, and cold starts

✅ Snapshot deployments are the fastest way to go from dev to prod without touching infra.

Iterative re-deploy

Once your deployment is live and working, you can simply go back to your Studio and start working on the next version of your API. Once ready to re-deploy, simply update the deployment.

How it works:

• Click on the Deployment Plugin and select the deployment you have already created
• Click "Update" Button on the top right.
• After reviewing the settings, click "Update" to validate the re-deployment

Lightning AI handles:

• Packaging a new version of the Studio environment
• Re-deploying the deployment gracefully using a rollout strategy
• Automatic revert if something goes wrong.

Revert versions

If you want to restore a previous version, simply find the release you want and restore it.

How it works:

• Click on the Deployment Plugin and select the deployment you have already created
• Click "Releases" tab
• Click on the 3 dots at the beginning of the release you want to restore
• Confirm the restoration of the previous release

Lightning AI handles:

• Keep the snapshot history
• Efficient and gracefully handles rollout

Via the SDK

Single server

By running this code within a Studio, it takes the current enviroment of the Studio and run it as a deployment

from lightning_sdk import Studio, Deployment
from lightning_sdk.deployment import HttpHealthCheck

deployment = Deployment()
deployment.start(
    studio=Studio(),
    command="python -m http.server",
    ports=[8000],
    health_check=HttpHealthCheck(
        path="/",
        port=8000
    )
)

In this other example, this code defines a FastAPI server with basic endpoints and orchestrates its deployment using the Lightning SDK's Studio and Deployment classes. It automates setting up a development environment, uploads the server code, and validates the deployment by sending a test request to ensure it returns the expected response.

from lightning_sdk import Studio, Deployment
from lightning_sdk.deployment import HttpHealthCheck
from datetime import datetime

server_code = """
from fastapi import FastAPI

app = FastAPI()

@app.get("/")
def read_root():
    return {"Hello": "World"}
"""

studio_name = "studio-" + datetime.now().strftime("%m-%d_%H:%M:%S")
studio = Studio(studio_name)
studio.start()

studio.run_with_exit_code("pip install 'fastapi[standard]'")

with open("server.py", "w") as f:
    f.write(server_code)
studio.upload_file("server.py")

# Create a Deployment from that Studio
deployment = Deployment()
deployment.start(
    studio=studio,
    command="fastapi run server.py",
    ports=[8000],
    health_check=HttpHealthCheck(
        path="/",
        port=8000
    )
)

resp = deployment.get("/")
assert resp.json() == {"Hello": "World"}

Multiple servers

This code utilizes the Lightning SDK to launch multiple Python server scripts, server_1.py and server_2.py, concurrently within a specified development environment labeled "your-studio-name". It sets up a deployment that makes both servers accessible on ports 3000 and 3001, respectively.

from lightning_sdk import Studio, Deployment

commands = [
    "python server_1.py --port 3000 &",
    "python server_2.py --port 3001 &",
]

deployment = Deployment()
deployment.start(
    studio=Studio("your-studio-name"),
    commands=commands,
    ports=[3000, 3001],
)

Update Secrets

This code updates an existing deployment by adding a secret and an environment variable using the Deployment class from the lightning_sdk. The update method adds a secret named "my-secret" and an environment variable "env-name" with the value "env-value" to the specified deployment configuration.

from lightning_sdk.deployment import Deployment, Secret, Env

deployment = Deployment("existing-deployment")
deployment.update(env=[Secret(name="my-secret"), Env(name="env-name", value="env-value")])

This code snippet updates the environment configuration of a target deployment using the environment variables from a source deployment. It initializes two deployment instances (source_deployment and target_deployment) and then copies the environment configuration (env) from the source_deployment to the target_deployment using the update method.

from lightning_sdk.deployment import Deployment, Secret, Env

source_deployment = Deployment("source-existing-deployment")
target_deployment = Deployment("target-existing-deployment")

target_deployment.update(env=source_deployment.env)