Audio cloaking and watermarking software

Several academic projects, platform-level consent systems, and detection or cloaking tools exist to combat unauthorized voice cloning, deepfakes, and adversarial voice scraping.engineering.washu.edu

Proactive Voice Cloaking & Adversarial Protection Tools

• AntiFake: Developed by researchers at Washington University in St. Louis (led by Professor Ning Zhang), AntiFake is a proactive defense tool that injects subtle perturbations into voice recordings. These adjustments disrupt how AI models extract acoustic features, preventing unauthorized text-to-speech engines from successfully cloning or mimicking a person's voice. The source code is publicly available for researchers and developers.engineering.washu.edu
• DeFake: A emerging framework and alternative research tool focused on disrupting deepfake generation pipelines and audio attribution.
• Stem Safe Solutions ("Empire Engine"): A specialized system aimed at music creators and vocalists to embed proof-of-ownership data into audio stems to resist unauthorized AI training and voice-to-voice re-synthesis.facebook.com

Audio Watermarking & Provenance Technologies

• Google SynthID (Audio): Developed by Google DeepMind, SynthID embeds an inaudible digital watermark directly into the audio waveform. The watermark survives through compression, speed changes, and analog-to-digital conversions, allowing platforms or detectors to verify whether an audio clip was generated by AI.
• ArtyShield / ArtyShield VoiceShield: Specialized audio protection and watermarking services designed to shield voice assets and creator media from unauthorized scraping and deepfake fabrication.

Platform-Side Consent & Speaker Verification

If your concern is legitimate platforms using your voice without authorization, major speech synthesis APIs enforce technical speaker-verification safeguards rather than relying solely on terms of service:

Detection & Consumer Alert Tools

• McAfee Deepfake Detector: A consumer-facing browser and application tool designed to analyze video and audio streams in real time to flag synthetic or cloned voices used in scams.

If you are trying to protect a specific type of audio (such as public speech recordings, music vocals, or private voice notes), let me know and I can detail how to implement these tools or safe recording practices.

There are tools designed to make your voice harder to clone without your consent, usually called audio cloaking, voice protection, or adversarial audio perturbation. There are also watermarking tools that help establish where an audio recording came from.

The key distinction is that these solve different problems: audio cloaking tries to interfere with cloning before you publish a recording; watermarking helps identify or trace audio afterward. Neither currently offers a guaranteed way to stop someone from copying your voice. The FTC also cautions that no single technical solution is sufficient.search.ftc.gov+1

1. Tools that can help protect your voice

• 

1. AntiFake / DeFake — Washington University in St. Louis

(Best starting point for personal recordings)

Adds subtle, adversarial changes to audio before you publish it, with the aim of making AI voice cloning less effective while preserving how the recording sounds to people. It's open-source research, not a guaranteed commercial-grade shield. Research has shown that attackers can sometimes remove the protection.untrainable.org+1

Access:AntiFake overview and resources

---

• 

2. Mask — getmask.org

Designed to protect voice recordings, music, and videos by subtly altering audio so AI systems have more difficulty analyzing or copying it.

Availability caveat: its website lists a waitlist, so confirm whether you can actually use it before relying on it.getmask.org

Website:getmask.org

---

• 

3. VoiceCloak — AAAI 2026 research

A newer defense targeting diffusion-based voice cloning. It modifies reference audio to interfere with speaker identity extraction and reduce the quality of unauthorized clones.

This is a research framework, not necessarily a ready-to-use consumer service.ojs.aaai.org

Paper:VoiceCloak at AAAI

2. Companies and tools for audio watermarking

These are more useful if your goal is to prove that a recording is authentic, trace synthetic content, or support a later dispute.

• 

AudioSeal — Meta AI

Embeds an imperceptible watermark in supported audio-generation workflows and provides a detector to locate marked AI-generated speech. It is open-source and developer-oriented.

Important: it does not cloak your original voice or prevent someone from cloning it.

Research and code ·GitHub

• 

Resemble AI

Offers audio watermarking, synthetic-media detection, and provenance technology. Its current products are particularly relevant to organizations that need to identify suspicious audio or verify media origins.www.resemble.ai

Website and products

• 

Evixar / SIGNED SOUND

A commercial provenance option combining audio watermarking, audio fingerprinting, and C2PA Content Credentials. More appropriate for publishers, media organizations, and content owners than for an individual trying to block cloning of their voice.www.evixar.com

Company and technology

3. Which approach fits your situation?

The most important caveat is that a watermark is not the same as consent enforcement. Someone can potentially clone your voice from an unprotected copy, remove a watermark, or use a system that ignores the protection. The FTC recommends a combination of technical safeguards, authentication, and platform-level protections.search.ftc.gov+1

One question so I can narrow this down

Where are you most concerned about your voice being copied?

1. Where is your voice exposed?

Social media and public videos

Podcasts, music, or professional recordings

Phone calls and live conversations

Private recordings or voice messages

---

2. What matters most?

Prevent cloning before it happens

Detect misuse and prove a recording is mine

Both prevention and proof

---

3. What kind of solution would you use?

An easy website or app

Developer tools or open-source software

A commercial service for a business or creator

[Find the best options for me]

Yes. There are tools designed to make it harder to clone your voice without permission, including audio cloaking, voice-protection filters, and audio watermarking. The important distinction is that some tools try to interfere with AI voice cloning, while others help you prove that an audio recording is authentic or trace its origin.

No current technique can guarantee that your voice cannot be cloned. The best approach depends on whether you're protecting recordings before you publish them, protecting live speech, or trying to detect and respond to unauthorized cloning.

Tools worth looking at

• 

1. Digital Cloak

(Easy starting point)

Best for: protecting recordings before posting them publicly.

Its website offers a browser-based tool that adds protective changes to audio, aiming to keep speech natural to human listeners while making AI cloning less reliable. It says processing happens in your browser. Test the output yourself before publishing.digitalcloak.app

Try Digital Cloak

---

• 

2. AntiFake / DeFake

Best for: adding anti-cloning perturbations to voice recordings.

This research-based approach introduces subtle audio changes intended to confuse voice-cloning systems without noticeably changing the recording to people. DeFake was recognized in the FTC's Voice Cloning Challenge. However, newer research indicates that some defenses can be bypassed, so treat this as a layer of protection, not a guarantee.search.ftc.gov+2

Read about AntiFake

---

• 

3. VoiceCloak

Best for: exploring newer anti-cloning research.

Published in the 2026 AAAI technical program, VoiceCloak uses several forms of audio and speaker-identity disruption to interfere with diffusion-based voice cloning. It is a research framework, so check the available code and demonstrations before expecting a simple consumer-ready app.ojs.aaai.org

Read the VoiceCloak research paper

---

• 

4. AudioWatermarking AWT2

Best for: tracing or authenticating recordings.

Embeds imperceptible identifying information in audio so a watermark can later be recovered. This is useful for professional audio distribution, licensing and tracking leaks, but it does not directly prevent someone from cloning your voice.audiowatermarking.com+1

Explore AWT2

Which kind of protection do you actually need?

The FTC specifically cautions that watermarking has limitations and that no single intervention solves voice cloning.www.ftc.gov

My recommendation

If you're a creator, podcaster, professional, or anyone who posts voice recordings online, I'd start by testing Digital Cloak on a copy of a recording before uploading it. Compare the original and protected versions for:

• Speech clarity and naturalness.
• Whether a cloning service can still reproduce your voice from the protected sample.
• Whether the protection survives common audio compression and editing.
• Whether the service processes audio locally or uploads it to a server.

Also, avoid publishing long, clean recordings of your voice unnecessarily, and use a separate verification method for sensitive requests such as money transfers. Don't rely on a voice sample alone to authenticate someone's identity.

One question so I can narrow this down

Where do you primarily want to protect your voice?

Voice recordings I post on social media, YouTube, or podcasts

Live speech, calls, or video meetings

Private recordings that I share with a limited audience

My voice has already been cloned or impersonated

---

What matters most?

Free and easy to use

Preserving my natural voice quality

The strongest practical anti-cloning protection

Keeping my recordings private and on my device

[Find the best options]