---
id: descript-audiotools
version: "0.7.2"
license: MIT
license_treatment: permissive
maintenance: abandoned
---

# descript-audiotools — Utilities for handling audio.

License: permissive · Maintenance: abandoned · Downloads: 464.6K/mo

## What it is and what it does

AudioTools is an object-oriented library for working with audio signals in Python. It wraps common audio operations—loading, filtering, playback, augmentation—into a unified API centered on the AudioSignal class, which represents an audio file or in-memory signal. The library handles batching, padding, and other preprocessing tasks commonly needed in audio machine learning pipelines.

The package depends on a large stack of scientific and audio libraries: numpy, scipy, librosa, torch, torchaudio, soundfile, and others. It is designed for developers building audio processing workflows, particularly in machine learning contexts. However, the package has been abandoned for over three years, meaning its dependencies may drift out of sync with current library versions, and bugs or compatibility issues are unlikely to be fixed.

Use it for:

- Load, slice, and apply audio filters (low-pass, high-pass) to WAV files for preprocessing training data.
- Batch and augment audio signals for machine learning model training with consistent padding and normalization.
- Integrate audio signal handling into research or educational projects that already depend on torch and librosa.
- Playback and inspect audio signals programmatically during development or debugging of audio pipelines.

## Worth the install?

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

Provides object-oriented audio signal handling with fast augmentation, batching, padding, and playback capabilities for audio processing workflows.

No. The package is abandoned (1122 days since last release) with no maintenance signal, and its 22 heavy dependencies—particularly torch and torchaudio—create a high risk of compatibility drift. Unless you are locked into an existing codebase already using it, prefer actively maintained audio libraries like librosa, soundfile, or scipy.signal for new projects.

## Install

pip install descript-audiotools
uv add descript-audiotools
poetry add descript-audiotools

## Installing descript-audiotools

Before you install: Low install friction with a pure-Python wheel, but the package is abandoned as of 1122 days since its last release. Its 22 runtime dependencies include heavy libraries like torch, torchaudio, and librosa, which may take time to install depending on your environment.

License in practice: MIT license is permissive, allowing use in commercial and private projects with minimal restrictions.

Quickstart:

pip install descript-audiotools

import audiotools
from audiotools import AudioSignal

signal = AudioSignal("audio.wav", offset=5, duration=5)
signal.low_pass(8000)

Requires torch and torchaudio, which have system-level dependencies and can take significant time to install; no explicit Python version requirement stated.

Verify before relying:

- Whether the package's 22 dependencies remain compatible with current versions of torch, scipy, librosa, and other heavy libraries given the 1122-day abandonment period.
- Whether the package works reliably with modern Python versions beyond 3.7, given no explicit requires_python constraint.

## Package facts

- License: MIT (permissive)
- Python support: unspecified
- Install friction: low
- Maintenance: abandoned
- Downloads: 464.6K/month (top 15,000 on PyPI)
- Known vulnerabilities: none known

## Tags

audio signal processing library, audio augmentation batching, wav file manipulation python, audio filtering and effects, audio data loading and preprocessing, audio-processing, abandoned-project

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