Davia connects your Python backend to a custom frontend — without requiring any frontend code. You write logic in Python, launch the server, and describe the UI you need in plain language. Davia takes care of the rest.

​
1. Set up your backend

Start by writing your backend logic using Python functions.

To make your app visible to Davia, you wrap your logic in a Davia instance and launch it with:

from davia import Davia, run_server

davia_app = Davia()

if __name__ == "__main__":
    run_server(davia_app)

​
2. Add tasks or graphs

You can expose your logic in two different ways:

  • @davia_app.task — for any standalone function (e.g. generate a report, summarize text, send a Slack message)
  • @davia_app.graph — for structured AI agents using LangGraph or CrewAI

Once defined, these functions become available for UI generation inside Davia.

​
3. Open the editor

When you run your script, Davia automatically starts a local server and opens the editor in your browser — already connected to your app.

🔗 The editor opens at:
https://dev.davia.ai/dashboard?entrypoint=http://127.0.0.1:2025

There’s nothing else to configure. Davia reads your tasks and graphs directly from your code.

​
4. Describe your interface

Inside the editor, you’ll see your available functions. Just describe the kind of app you want.

For example:

Build a simple dashboard with two buttons: one to generate a daily report using generate_report, and another to email it using send_email.

Davia creates the corresponding interface, wired to your backend. You can test it immediately and iterate on the logic or layout at any time.

​
5. Iterate visually, code freely

You stay in control of your backend in Python. Davia gives you full flexibility on the logic — while letting you skip all the time spent on frontend scaffolding, wiring, and state management.