---
id: pooch
version: "1.9.0"
license: BSD-3-Clause
license_treatment: permissive
maintenance: active
---

# pooch — A friend to fetch your data files

License: permissive · Maintenance: active · Downloads: 17.0M/mo

## What it is and what it does

Pooch is a Python library that handles the tedious parts of downloading and caching data files. It sits between your code and the network, fetching files only when needed, storing them locally, and verifying they haven't been corrupted or replaced. It supports HTTP, FTP, and DOI-based downloads from repositories like Zenodo and figshare, with built-in hash verification to ensure reproducibility.

For scientists and researchers, Pooch automates the distribution of sample datasets—you can host data on a repository and let your package download it automatically, ensuring everyone using your code has the same version of the data. For package developers, it provides a clean API to manage versioned data caches without forcing users to manually download files or deal with urllib and requests directly.

Use it for:

- Download sample datasets for documentation or tutorials without embedding large files in your package repository.
- Distribute versioned data files with your Python package, automatically fetching them on first use.
- Verify downloaded files against known hashes to detect corruption or tampering.
- Host data on Zenodo or figshare and reference it by DOI instead of managing URLs.
- Implement reproducible research workflows where colleagues run the same code and get identical data.
- Cache remote data locally to avoid re-downloading during repeated script runs or development iterations.

## Worth the install?

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

Pooch downloads files from HTTP, FTP, and data repositories (Zenodo, figshare), caches them locally, and verifies integrity via hash checking—eliminating manual file management and urllib boilerplate.

Yes. Pooch is production-stable (Development Status 5), actively maintained, has no known vulnerabilities, and is already trusted by major scientific Python projects (SciPy, scikit-image, xarray, napari, and others). Install it if you need to download, cache, and verify data files in any Python project—it eliminates boilerplate and adds reproducibility with minimal overhead.

## Install

pip install pooch
uv add pooch
poetry add pooch

## Installing pooch

Before you install: Low friction: pure Python with only three runtime dependencies (platformdirs, packaging, requests). Active maintenance with last commit 2026-08-04 and no known vulnerabilities.

License in practice: BSD-3-Clause is permissive; you can use, modify, and distribute Pooch with minimal restrictions, making it suitable for both open-source and proprietary projects.

Quickstart:

import pooch

# Download and cache a file, verifying its hash
fname = pooch.retrieve(
    url="https://example.com/data.csv",
    known_hash="md5:a7332aa6e69c77d49d7fb54b764caa82"
)

# For package developers, create a managed cache
GOODBOY = pooch.create(
    path=pooch.os_cache("mypackage"),
    base_url="https://github.com/myproject/data/",
    registry={"file.csv": "sha256hash"}
)
fname = GOODBOY.fetch("file.csv")

Verify before relying:

- Whether custom post-processors (unzip/decompress) are included in the base package or require additional dependencies.
- Performance characteristics when managing large numbers of files or very large individual files.
- Whether FTP support is fully tested and maintained or primarily HTTP-focused.

## Package facts

- License: BSD-3-Clause (permissive)
- Python support: supports_current
- Install friction: low
- Maintenance: active
- Downloads: 17.0M/month (top 5,000 on PyPI)
- Known vulnerabilities: none known

## Tags

file download caching, data file management, http download with cache, sample dataset distribution, file integrity verification, zenodo figshare download, local data cache, reproducible data fetching, data-distribution, reproducible-research, caching

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