Deployment

You can host a Datapane app from inside your current Python environment, whether that's your local machine, or a hosted notebook like Colab or Databricks.

Local App Serving¶

To host your app, run dp.serve_app.

Note

A port will be chosen automatically, but you can manually specify port and host.

import datapane as dp

view = dp.View(...)

dp.serve_app(view)

Creating a public URL¶

Datapane has built-in support for ngrok, which allows you to create a public, shareable URL for your data app. Simply add public=True to your serve command. This will prompt you to enter your ngrok token, and will return a URL which you can share.

dp.serve_app(app, public=True)

Embed mode¶

If you are embedding your app in a third-party product, you will need to pass embed_mode=True to dp.serve. This disables specific security policies which will not work within an iframe.

Datapane Cloud¶

Datapane Cloud will shortly provide App hosting specific and optimized to the needs of App running and hosting, including access-controls, security and authentication features that organizations require.

At the same time, you can easily host your app via any web server hosting service (especially those that work with Python / WSGI-based apps), including the services listed here.

Fly.io¶

Fly.io provides a simple way to build and deploy your app. Fly has a generous free tier where you can host your data app on the web. Note that the free tier is limited to 256mb of RAM, so if you are doing any data processing, you may want to upgrade.

Getting started¶

You will need:

One of either:

A requirements.txt for pip dependencies
A Dockerfile

Datapane CLI provides a helper for generating a Dockerfile:

$ datapane app generate dockerfile

Deploying your app¶

Once done, you can deploy your app with the following commands:

Create and configure a new app from source code or a Docker image.
```
$ flyctl launch --remote-only --build-only
```

Deploy your app

$ flyctl deploy --remote-only -e 'PORT=8080'

Save your config file for future deploys
```
$ flyctl config save
```

This will:

Generate a fly.toml file
Push and build your app on their servers
Provide you with a URL to access your App

Render.com¶

Render provides a simple way to build and deploy your projects from a git repository.

To get started, you'll need:

A Render account
Your source code in a git repository
A requirements.txt for pip dependencies
A Dockerfile. You can generate this with datapane app generate dockerfile

Once done, you can setup a deployment from the dashboard:

Create a new Web Service
Select your repository
Name your service
click Create (at the end of the page)

This will start building your App, then deploying it. Once complete, you'll be able to find the URL in the Dashboard.