Resemble AI

What is Resemble AI?

Resemble AI offers deepfake detection for audio, video and images, plus watermarking for provenance and C2PA compliance. Its product lines include Detect, Intelligence (explainable results), PerTh multimodal watermarking, Identity (voice and likeness enrollment), Meetings protection and a Chrome extension. It targets enterprises in fraud, security and trust and safety, along with developers.

Key capabilities of Resemble AI

• Detect: real-time deepfake detection; the vendor claims a real-time factor under 1.0 for its audio model.
• Explainable results: human-readable explanations of why media was flagged.
• Single REST API: one API covers audio, video and image detection, with a Python SDK.
• Meeting protection: detection in Microsoft Teams, Zoom, Google Meet and Webex.
• Watermarking: PerTh multimodal watermarking for provenance.
• Identity: enrollment of voices and likenesses.
• Flexible deployment: includes zero-retention mode and air-gapped on-premises deployment.

How Resemble AI works

You send media to the REST API or connect Resemble to a contact center or meeting platform. It scores the content in real time and returns a result with an explanation. In zero-retention mode, media is deleted after detection. The vendor reports accuracy of 99.5% for audio, 98.2% for video and 95.8% for images, which are self-reported figures.

Who uses Resemble AI?

Enterprises in fraud and identity, security and risk, and trust and safety, including contact centers, telecoms, finance, media, healthcare and the public sector, plus developers building detection into products.

Resemble AI pricing

The vendor offers a Start free option and links to a pricing page, but does not publish specific prices on its main page. Custom enterprise deployment is available through the vendor.

Resemble AI alternatives

Respeecher is known for voice conversion and synthetic voice production, so it sits on the voice creation side rather than detection. Speechmatics is a speech recognition provider and does not focus on deepfake detection. Hume AI builds expressive voice and emotion AI for interfaces.