Resemble AI underwent a major strategic shift in August 2026: it stopped selling voice generation to new customers and pivoted entirely to AI deepfake detection and content provenance. The current product line is built around security: Detect performs real-time detection of AI-generated audio, video, and images, with the vendor claiming 99.5% audio detection accuracy based on an August 2026 Podonos benchmark; Intelligence provides a readable explanation for each verdict; Watermarker embeds imperceptible watermarks into audio, video, and images with C2PA provenance records; Identity registers a speaker from 4 seconds of audio and uses it to verify or flag impersonation; Meetings detects synthetic voice and face in Zoom, Teams, Meet, and Webex in real time without adding a bot; Agent Detection identifies AI agents. Customers listed on the site include Netflix, Paramount, Deutsche Telekom, Five9, and 8x8. Security posture covers SOC 2 Type II, ISO 27001, GDPR, and HIPAA alignment, with on-premise and air-gapped deployment options. Pricing is public: Flex $0/month (pay-as-you-go credits, no subscription), Team $350/month (annual $280), Business $1,000/month (annual $800), Enterprise custom. Important clarification: Resemble was previously known for voice cloning and TTS, but new customers can no longer purchase voice generation—older TTS or cloning plans found in search results are historical and no longer represent the current product. Its open-source speech models (such as Chatterbox) remain freely available under the Voice AI Research line but are no longer sold commercially.

Pros & Cons:
Pros:
The post-pivot product line has a coherent logic—Detect + Intelligence + Watermarker + Identity + Meetings form a complete enterprise security loop of detection, explanation, provenance, and verification; the claimed 99.5% audio detection accuracy has a Podonos benchmark behind it, and independent evaluations list it among the leaders in audio deepfake detection; Meetings detects synthetic media without injecting a bot into the call, which is a concrete design for protecting high-risk individuals such as C-suite executives; SOC 2 Type II, ISO 27001, HIPAA alignment, and on-premise/air-gapped options give it real appeal for regulated industries such as finance, healthcare, and government; the free Flex tier with pay-as-you-go credits that never expire lowers the barrier for enterprise evaluation.
Cons: Voice generation/TTS is no longer available to new customers—if your need is voice cloning or speech generation, this product is not for you now, and you would need to look at its open-source non-commercial models or alternatives; independent third-party reviews are limited, and the 3.9/5 G2 rating (20 reviews) aggregated by Techreviewer reflects the old voice product line, with complaints about naturalness and pronunciation that are unrelated to the current detection product but easy to confuse; detection accuracy figures come mainly from vendor claims or specific benchmarks, with no large-scale independent cross-validation yet; pricing is high for small teams (Team starts at $280/month annual), and detection is billed per second/image/run, so cost needs careful modeling at volume.

Who is it for?:Enterprise security, risk, and compliance teams—especially in regulated industries such as finance, healthcare, and government—that need to detect AI-generated voice or face in meetings, support recordings, or uploaded media; content owners that need to watermark their own audio/video and attach C2PA provenance records; existing Resemble voice-model customers who continue to receive support (though new customers cannot buy voice generation). Not suited for new users who need voice cloning or TTS generation, or budget-constrained individuals looking for a simple detection tool.