Skip to content
SPECIFICTION
Get listed

DeepFilterNet

github.com

Free plan Open source WindowsmacOSLinuxAPISelf-hosted
Visit github.com
Spec score 8.9 #1 of 65 in Software Composition Analysis Software

What it is

DeepFilterNet is a free, open-source framework for speech enhancement on full-band 48 kHz audio, using deep filtering. Its command-line tool suppresses noise in WAV files, with 48 kHz as its only supported sampling rate. A LADSPA plugin provides real-time noise suppression and can be used with PipeWire filter chains to create a virtual microphone or audio output, including for VoIP apps such as Zoom or Discord. The plugin's minimum latency is 20 ms from STFT processing, plus any delay added by the LADSPA host. The Rust command-line tool defaults to DeepFilterNet2; the Python interface documents DeepFilterNet and DeepFilterNet2. An optional post-filter may slightly over-attenuate very noisy sections. The framework supports Linux, macOS and Windows, though training is tested only under Linux. The Python package installs as deepfilternet, and an optional train extra adds data-loading functionality for Linux training. Code is available under either the MIT License or Apache License 2.0.

Spec sheet

What it does
DeepFilterNet is a low-complexity speech-enhancement framework that uses deep filtering for full-band audio.Read 15 Oct 2026 · github.com
Purpose
DeepFilterNet is a low-complexity speech enhancement framework for full-band 48 kHz audio using deep filtering.Read 4 Oct 2026 · github.com
Security
The repository page shows a Security and quality tab but provides no security or compliance claims in the project information read.Read 4 Oct 2026 · github.com
License
The repository code is free and open source under either the MIT License or Apache License 2.0, at the user's option.Read 4 Oct 2026 · github.com
Audio coverage
The framework targets 48 kHz full-band speech audio.Read 15 Oct 2026 · github.com
CLI limitation
The precompiled deep-filter command currently accepts only WAV files sampled at 48 kHz.Read 15 Oct 2026 · github.com
Distribution
The project provides precompiled binaries through its GitHub releases and a Python wheel through PyPI.Read 4 Oct 2026 · github.com
Embedded performance
The DeepFilterNet2 paper reports a real-time factor of 0.04 on a notebook Core-i5 CPU and describes embedded real-time use.Read 15 Oct 2026 · arxiv.org
Framework platforms
The framework supports Linux, macOS, and Windows; training is only tested under Linux.Read 4 Oct 2026 · github.com
Hugging Face demos
Hendrik Schröter's Hugging Face profile lists two DeepFilterNet Spaces, including one marked Runtime error.Read 15 Oct 2026 · huggingface.co
Latency
The LADSPA plugin has a minimum latency of 20 ms from STFT processing, plus latency from the LADSPA host.Read 4 Oct 2026 · github.com
Latest release
GitHub lists version 0.5.6 as the latest release and says it fixes issues 421 and 293.Read 15 Oct 2026 · github.com
Licence
The repository is free and dual-licensed so users may choose MIT or Apache License 2.0.Read 15 Oct 2026 · github.com
Limit: ASR use
A Hacker News commenter cautioned that DeepFilterNet would not improve an automatic-speech-recognition model such as Whisper.Read 15 Oct 2026 · news.ycombinator.com
Limit: Colab installation
A 2025 GitHub issue reports metadata-generation failure while installing deepfilternet 0.5.6 in Google Colab.Read 15 Oct 2026 · github.com
Limit: dependency setup
A Hacker News user reported frustrating Poetry setup involving keyring behavior over SSH.Read 15 Oct 2026 · news.ycombinator.com
Limit: package installation
A GitHub issue reports that installing deepfilternet 0.5.3 failed because the required deepfilterlib 0.5.3 distribution was unavailable.Read 15 Oct 2026 · github.com
Models and processing
The Rust command-line tool defaults to DeepFilterNet2, while the Python interface also documents use of DeepFilterNet and DeepFilterNet2 models.Read 4 Oct 2026 · github.com
Noise suppression
Its command-line tool suppresses noise in WAV recordings.Read 15 Oct 2026 · github.com
Package activity
SkillFed lists the package as dormant, with its latest release dated 2023-08-31.Read 15 Oct 2026 · skillfed.io
Package downloads
SkillFed reports 78,622 monthly downloads for deepfilternet in its displayed 30-day window.Read 15 Oct 2026 · skillfed.io
PipeWire integration
The LADSPA plugin can be used in PipeWire filter chains to create a virtual noise-suppressed microphone.Read 15 Oct 2026 · github.com
Post-filter
An optional post-filter slightly over-attenuates very noisy sections.Read 4 Oct 2026 · github.com
Postfilter
An optional postfilter is documented as slightly over-attenuating very noisy sections.Read 15 Oct 2026 · github.com
Praise: CPU efficiency
A Hacker News commenter reported adequate inference speed on plain CPU using an ONNX-based deployment.Read 15 Oct 2026 · news.ycombinator.com
Praise: overall capability
A Hacker News commenter called DeepFilterNet the most well-rounded open-source AI noise-suppression tool they had found.Read 15 Oct 2026 · news.ycombinator.com
Pretrained models
The command-line and Python workflows can load pretrained DeepFilterNet, DeepFilterNet2 and DeepFilterNet3 models.Read 15 Oct 2026 · github.com
PyPI maintainer
PyPI identifies Rikorose as the single maintainer and Hendrik Schröter as author.Read 15 Oct 2026 · pypi.org
PyPI runtime requirement
The PyPI package declares Python 3.8 or newer and below Python 4.0.Read 15 Oct 2026 · pypi.org
Python API
Applications can call the init_df and enhance functions to load a model and process noisy audio.Read 15 Oct 2026 · github.com
Python distribution
DeepFilterNet is installable from PyPI with pip, and an optional train extra adds data-loading functionality.Read 15 Oct 2026 · github.com
Python installation
The Python package installs with pip as deepfilternet, and an optional train extra adds data-loading functionality for training on Linux.Read 4 Oct 2026 · github.com
Real-time plugin
The project includes a LADSPA plugin for real-time noise suppression.Read 4 Oct 2026 · github.com
Real-time processing
The project provides a real-time Rust implementation and a LADSPA noise-reduction plugin.Read 15 Oct 2026 · github.com
Repository popularity
GitHub shows the repository with about 4.8 thousand stars and 534 forks.Read 15 Oct 2026 · github.com
Research performance
The 2023 paper reports a real-time factor of 0.19 on a single-threaded notebook CPU while matching state-of-the-art speech-enhancement benchmarks.Read 15 Oct 2026 · arxiv.org
Rust components
The repository contains Rust libraries for STFT/ISTFT processing, data loading and augmentation.Read 15 Oct 2026 · github.com
Supported operating systems
The framework supports Linux, macOS and Windows, while training is tested only on Linux.Read 15 Oct 2026 · github.com
Training
Training expects datasets in HDF5 format, with a dataset configuration file for train, validation, and test sets.Read 4 Oct 2026 · github.com
Training workflow
Training expects HDF5 speech, noise and room-impulse-response datasets plus a configuration covering train, validation and test sets.Read 15 Oct 2026 · github.com

Plans and prices

DeepFilterNet Free Free and open source · MIT or Apache-2.0 license Read 4 Oct 2026 · github.com
Free and open source Free MIT or Apache-2.0 licensing · self-hosted framework · Linux, macOS and Windows support Read 15 Oct 2026 · github.com

Every DeepFilterNet plan and price →

Strong points and limits

Strong points

  • Free and open source under MIT or Apache 2.0.
  • LADSPA plugin supports real-time noise suppression.
  • PipeWire filter chains can create a virtual microphone.
  • Framework supports Linux, macOS and Windows.

Limits

  • WAV processing currently supports only 48 kHz.
  • Training is tested only under Linux.
  • Plugin latency starts at 20 ms before host delay.
  • Post-filter may over-attenuate very noisy sections.

Who it suits

DeepFilterNet suits people seeking open-source noise suppression for 48 kHz audio, including real-time use through a LADSPA plugin. It also suits developers who want a Python package or command-line tool.

Verdict

DeepFilterNet offers command-line and real-time noise suppression, with open-source code and support across three desktop operating systems. Its 48 kHz limit and Linux-only training testing matter for prospective users.

Where Specifiction places it

Alternatives to DeepFilterNetAll alternatives →