---
id: neptune
version: "1.14.0.post2"
license: Apache-2.0
license_treatment: permissive
maintenance: abandoned
---

# neptune — Neptune Client

License: permissive · Maintenance: abandoned · Downloads: 144.6K/mo

## What it is and what it does

Neptune is a client library for the Neptune MLOps platform, designed to integrate with machine learning workflows to track experiments, log metrics and artifacts, and manage model metadata. It provides integration points for logging training runs, model parameters, and results to a centralized Neptune app instance, enabling teams to organize and compare ML work across projects.

The package depends on a broad set of utilities including requests, pandas, boto3, and websocket-client for cloud storage integration, OAuth authentication, and real-time communication. It supports Python 3.8 through 3.14 on multiple operating systems. The fact sheet indicates this is the legacy client for Neptune app version 2.x; the project repository is now archived, and users are directed toward neptune-client-scale for new work.

Use it for:

- Log and track ML experiment metrics, hyperparameters, and model performance during training runs.
- Store and retrieve model artifacts and training outputs in a centralized Neptune workspace.
- Compare multiple experiment runs side-by-side to identify the best model configurations.
- Integrate experiment tracking into existing ML pipelines and CI/CD workflows.
- Manage team collaboration on ML projects with shared experiment history and metadata.

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

Neptune is a Python client for experiment tracking and ML metadata management, allowing you to log, organize, and monitor machine learning workflows and model artifacts.

No. The repository is archived and the package is abandoned. While it remains at production/stable status and has no known vulnerabilities, the fact sheet explicitly directs users to neptune-client-scale for new work. Installing this package locks you into a legacy client with no active maintenance or security updates.

## Install

pip install neptune
uv add neptune
poetry add neptune

## Installing neptune

Before you install: Low install friction with a pure-wheel distribution. However, the repository is archived and marked as abandoned; the package is no longer actively maintained, though it remains at production/stable status.

License in practice: Licensed under Apache-2.0 (permissive), allowing broad use, modification, and distribution with minimal restrictions.

Quickstart:

pip install neptune

import neptune
run = neptune.init_run()

Requires Python 3.8 or later; requires credentials and network access to a Neptune app instance.

Verify before relying:

- Whether Neptune 2.x is still compatible with current Neptune app infrastructure or if migration to neptune-client-scale is required for ongoing use.
- Support status and security patch availability given the archived repository status.

## Package facts

- License: Apache-2.0 (permissive)
- Python support: supports_current
- Install friction: low
- Maintenance: abandoned
- Downloads: 144.6K/month (top 15,000 on PyPI)
- Known vulnerabilities: none known

## Tags

ML experiment tracking, machine learning metadata store, model registry python, experiment logging MLOps, ML model tracking client, mlops, experiment-tracking, legacy

[View on SkillFed](https://skillfed.io/packages/neptune) · [View on PyPI](https://pypi.org/project/neptune/)