---
id: neptune-query
version: "1.14.1"
license: Apache-2.0
license_treatment: permissive
maintenance: abandoned
---

# neptune-query — Neptune Query is a Python library for retrieving data from Neptune.

License: permissive · Maintenance: abandoned · Downloads: 1.4M/mo

## What it is and what it does

neptune_query is a read-only Python client that retrieves experiment metadata, run information, metrics, and logged attributes. It provides functions to list experiments and runs, fetch metadata as pandas DataFrames, download files, and define filters to query specific subsets of data. The package depends on httpx for HTTP requests, pandas for tabular output, protobuf for serialization, PyJWT for authentication, attrs, azure-storage-blob, python-dateutil, and tqdm.

The package is designed for data scientists and ML engineers who need to programmatically access project data for analysis, reporting, or downstream processing. It supports both experiment-level and individual run-level queries, with filtering capabilities based on attributes like ownership, archival status, and metric thresholds. However, the repository is archived and abandoned, with no commits since 2026-04-20, meaning no active maintenance or support is available.

Use it for:

- Export experiment metadata and metrics to pandas DataFrames for offline analysis or reporting.
- Programmatically filter and retrieve runs matching specific criteria using the filters module.
- Download files logged during experiments for post-hoc inspection or model evaluation.
- Query individual runs by ID instead of experiment names for fine-grained data access.
- Build automated pipelines that fetch data for integration with other tools.

## Worth the install?

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

Read-only API for querying experiment metadata, runs, metrics, and attributes with filtering and export to pandas DataFrames.

Yes, if you need read-only programmatic access to experiment data and can tolerate the abandoned maintenance status. Low install friction and permissive Apache 2.0 license make integration straightforward. However, no bug fixes or compatibility updates are forthcoming—install only if your use case is stable and unlikely to require updates, or if you can manage potential breakage with future Python or API changes.

## Install

pip install neptune-query
uv add neptune-query
poetry add neptune-query

## Installing neptune-query

Before you install: Low install friction with a pure-Python wheel. However, the package is abandoned as of 2026-04-20 with no active maintenance, which poses a risk for bug fixes or compatibility updates with future Python or dependency versions.

License in practice: Apache License 2.0 is permissive and poses no restrictions on commercial or private use. You may use, modify, and distribute freely provided you include the license notice.

Quickstart:

pip install 'neptune-query<2.0.0'

import neptune_query as nq

# Set environment variables first:
# export NEPTUNE_API_TOKEN="your-token"
# export NEPTUNE_PROJECT="workspace/project"

nq.fetch_experiments_table(attributes=["metrics/accuracy"])

Requires NEPTUNE_API_TOKEN and NEPTUNE_PROJECT environment variables to be set before use.

Verify before relying:

- Whether the abandoned status affects real-world usability or if the API remains stable against backend changes.
- Compatibility with current API versions or whether breaking changes have occurred since last release on 2026-04-07.

## Package facts

- License: Apache-2.0 (permissive)
- Python support: supports_current
- Install friction: low
- Maintenance: abandoned
- Downloads: 1.4M/month (top 5,000 on PyPI)
- Known vulnerabilities: none known

## Tags

experiment metadata query api, ml experiment tracking retrieval, fetch run metrics dataframe, mlops metadata access, experiment filtering and export, mlops, experiment-tracking, data-export

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