LIVE
Google Introduces "Gemini 4 Carbon" as Next Step in Its Model Lineup28/10/26 · Google|DeepMind unveils Genie 4, a new generation of world model27/10/26 · Google DeepMind|Anthropic files IPO prospectus, with risk disclosures taking center stage26/10/26 · Anthropic|Robot policy trained only in simulation matches top real-world systems26/10/26|OpenAI unveils GPT-6.1 Terra, a model built for computer use26/10/26 · OpenAI|Google's AI gives evasive answers about who won the 2020 US election28/10/26 · Google|NVIDIA introduces Nemotron 3.5 Transcribe, a promptable multilingual transcription model28/10/26 · NVIDIA|OpenAI and ServiceNow deepen partnership on enterprise workflows28/10/26 · OpenAI|Failed startups' data finds a second life training AI models28/10/26|Anthropic expands AI for Science program to 20 foundations and research institutions28/10/26 · Anthropic|Anthropic shares early results on using Claude to accelerate drug development27/10/26 · Anthropic|GEPA-Curator: Filtering Failures to Better Train LLM Agents27/10/26|Google Introduces "Gemini 4 Carbon" as Next Step in Its Model Lineup28/10/26 · Google|DeepMind unveils Genie 4, a new generation of world model27/10/26 · Google DeepMind|Anthropic files IPO prospectus, with risk disclosures taking center stage26/10/26 · Anthropic|Robot policy trained only in simulation matches top real-world systems26/10/26|OpenAI unveils GPT-6.1 Terra, a model built for computer use26/10/26 · OpenAI|Google's AI gives evasive answers about who won the 2020 US election28/10/26 · Google|NVIDIA introduces Nemotron 3.5 Transcribe, a promptable multilingual transcription model28/10/26 · NVIDIA|OpenAI and ServiceNow deepen partnership on enterprise workflows28/10/26 · OpenAI|Failed startups' data finds a second life training AI models28/10/26|Anthropic expands AI for Science program to 20 foundations and research institutions28/10/26 · Anthropic|Anthropic shares early results on using Claude to accelerate drug development27/10/26 · Anthropic|GEPA-Curator: Filtering Failures to Better Train LLM Agents27/10/26|
AppWatch outofficial publisher

DeepFilterNet

App : A product you use directly

Real-time noise suppression, lightweight, works even on CPU.

The most effective open-source denoiser for voice.

Official site GitHub · 4,882 ★
Overall score
6.6 /10
Security: nothing to flag
Utility
8.0
Reliability
1.0
Security
No caveat on record
Ease
Technical · Technical

Security reflects where data goes, requested permissions and incidents on record at the verification date. It is not a guarantee of zero risk: check the deployment mode and the connections you enable.

What it does, concretely

Audio recordings often contain unwanted background noise: fans, keyboards, street sounds or office ambience. DeepFilterNet addresses this problem by suppressing such noise in real time, without sending your files to a remote server.

The tool is based on a technique called Deep Filtering applied to full-band audio at 48 kHz. It comes in three successive versions (DeepFilterNet, DeepFilterNet2, DeepFilterNet3), each improving perceptual quality and reducing computational complexity. It can be used either via a precompiled command-line binary or via a LADSPA plugin integrated with PipeWire for live microphone filtering.

DeepFilterNet is aimed at developers, podcasters, streamers and advanced users on Linux or Windows who want to improve voice quality without relying on an online service. It can also be integrated into Python scripts through the package available on PyPI.

However, it is not suited for users looking for a simple graphical interface or working exclusively on macOS: the interactive demo and model training are only tested on Linux, and the tool remains command-line oriented.

What it offers

  • ·Real-time noise suppression on 48 kHz WAV audio files via the deep-filter command.
  • ·LADSPA plugin compatible with PipeWire for live microphone noise filtering during calls or recordings.
  • ·Three pretrained models available (DeepFilterNet, DeepFilterNet2, DeepFilterNet3) offering progressive quality levels.
  • ·Precompiled binary with no Python dependency, ready to use after download on Linux and Windows.
  • ·Python integration via pip (deepfilternet package) for use in scripts or audio processing pipelines.
  • ·Post-filter option (--pf) to further attenuate very noisy sections.
  • ·CPU processing without requiring a GPU, with optional GPU support via the PyTorch backend.

In practice

Price
Free
Commercial use
Commercial use: yes
Available on
cli, windows, linux
Licence
MIT
Runs locally
Yes
FR interface
No

Security and data

Rating
Security: nothing to flag
Your data
Data stays local
Maintenance
Inactive for over six months
Last release
31 August 2023

Links and repository checked on 25 October 2026 · added on 1 January 2021

How to install DeepFilterNet

1. Go to the GitHub releases page (github.com/Rikorose/DeepFilterNet/releases) and download the precompiled binary for your system (Linux or Windows). 2. Verify the integrity of the downloaded file if checksums are provided on the release page. 3. Place the binary in a folder accessible from your terminal (for example /usr/local/bin on Linux). 4. For Python installation, run: pip install deepfilternet (after installing PyTorch from pytorch.org). 5. Test the installation by running: deep-filter --version or deepFilter path/to/noisy_audio.wav. 6. All processing is done locally: no audio data is sent to an external service.
pip install deepfilternet

In the same category

Whole category Audio, voice & music →