Skip to main content
By the end you will have a deployed service reachable at https://<endpoint-id>.svc.uniac.ai.

​
Prerequisites

  • Python 3.11 or newer
  • Node 16 or newer (to install the CLI via npm)
  • Docker (for uniac dev)

​
1. Install the CLI and SDK

npm i -g @uniac/cli
pip install uniac

​
2. Log in

uniac auth login
This opens a browser to complete the sign-in flow and writes a token to ~/.uniac/auth.json.

​
3. Scaffold a project

uniac init
cd my-first-service
uniac init is interactive — pick app for the kind, Python for the language, and accept the default entrypoint. The result is a uniac.json manifest and a starter app.py. Replace app.py with a minimal System:
app.py
from uniac import System, load

class MyFirstService(System):
    def __init__(self, greeting: str = "hello"):
        self.greeting = greeting

service = load(MyFirstService)

# Your handler — Uniac runs whatever your entrypoint command starts.
# For a FastAPI app:
from fastapi import FastAPI
api = FastAPI()

@api.get("/hello")
def hello(name: str = "world"):
    return {"message": f"{service.greeting}, {name}!"}
Add a requirements.txt:
requirements.txt
fastapi
uvicorn[standard]

​
4. Run it locally

uniac dev
The simulator starts a local Docker stack and opens the dashboard. Send GET /hello?name=cloud from the dashboard’s request panel.

​
5. Deploy

uniac link
uniac deploy
uniac link binds the project to a remote Uniac project (created on the fly if it doesn’t exist). uniac deploy builds, pushes the image, and registers the deployment. The CLI prints the public URL on success: 
Deployed: https://ab12cd.svc.uniac.ai
Call it: 
curl https://ab12cd.svc.uniac.ai/hello?name=cloud

​
Next steps

Your first service

The same flow with deeper explanations.

Concepts

Services, Systems, Nodes, projects, deployments.