Skip to content

Using Neptune#

This tab contains various how-tos and instructions for using the Neptune API and web app.

We generally assume that you have the latest version of the client library installed.


  1. Register at
  2. Install the Neptune client library: pip install -U neptune (or see Install Neptune)
  3. Save your API token and project name as environment variables. (1)

You're good to go:

import neptune

run = neptune.init_run()
run["your/structure"] = some_metadata

# more model training code and metadata logging

  1. We don't recommend it, but you can also pass your info when initializing Neptune:

    import neptune
    run = neptune.init_run(

Examples on GitHub#

In the neptune-ai/examples repo on GitHub, you can browse our full library of example scripts, notebooks, use cases, and Neptune projects.

View code examples 

Neptune example project#

To get an idea of what a Neptune project can be used for, you can browse an example project set up for image segmentation.

Tip: Take a look at the script in the Source code view to see how the metadata was logged.

Browse example project 

See also#

Use cases Comprehensive, practical tutorials

Integrations Use Neptune with a specific framework or library

Management Invite people, set permissions, edit project settings, or change your plan