3 billion parameters validated against 160+ generative AI tools. Rather than matching against a library of known fakes, DETECT-3B Omni learns the statistical artifacts that generative architectures leave behind. Delivered seamlessly via an automated security bot that joins your calls.
Detection accuracy across audio, video, and images under real-world attack conditions.
Benchmarks available upon request.
RESEMBLE meetings attack coverage
Can it catch a fake mid-call? Every time.
Face swap and AI filter impersonation
DETECTED
DETECTED
Replay attack
DETECTED
DETECTED
Voice cloning
DETECTED
DETECTED
Partial manipulation (localized face or voice edit)
DETECTED
DETECTED
Synthetic video persona (deepfake candidate or executive)
DETECTED
DETECTED
Live call on Zoom, Teams, Google Meet, or Webex
COVERED
COVERED
OUR CAPABILITIES
Automate detection — no set-up per call
We pride ourselves on our ability to challenge core assumptions, unpick legacy behaviors, streamline complex processes, and strike a balance between great design and functional development.
Bot joins automatically
Syncs with your organization's calendar and joins every scheduled meeting automatically, with no manual triggers and no action required from participants or candidates.
Alerts in real time
Notifies security teams instantly via email, Slack, or SMS the moment a deepfake is identified, with a confidence score and recommended action.
Forensic evidence on every detection
Generates frame-by-frame analysis with timestamp data, manipulation technique identification, and certainty scoring. Exports reports for regulators, legal counsel, and executive briefings.
99%
Accuracy on the latest video generation models incl. Veo 2
4
Platforms covered: Zoom, Teams, Meet, and Webex
51
Languages covered with no enrollment required
AI AttacK riSK BY teams
A deepfake call only needs a few seconds
The exposure is different for every team. The call is the same.
Security & Finance
Industry risk:
Executive impersonation on live calls authorizing wire transfers or credential handovers.
Resemble AI solution:
Alert the security team before the transfer is approved or your IP is leaked.
HR & Talent
Industry risk:
Deepfake candidates using face swaps to pass recruiter screens undetected.
Resemble AI solution:
Recruiters are notified in-call, before the candidate advances.
Security Operations
Industry risk:
Live meetings are the most direct attack surface with no existing detection coverage.
Resemble AI solution:
Every scheduled meeting across all four platforms is covered automatically.
Legal & Compliance
Industry risk:
No documented evidence when fraud is reported after the call ends.
Resemble AI solution:
Every detection produces a forensic report exportable for regulators and counsel.
By default, yes. Detect 2B appears as a named participant. The bot name is fully customizable. Some organizations display it openly as a detection tool, others name it as a note taker.
Does the bot record the meeting?
No. Detection runs in real time against live audio and video. No call content is stored. The only output is the detection verdict, confidence score, and forensic report. This simplifies GDPR and consent obligations significantly.
How does calendar integration work?
Connect your Google Calendar or Outlook via the dashboard. Earmark specific users or filter by meeting type: all meetings, external-only, or host-only. The bot joins automatically from that point forward. No action required per call.
What happens when a deepfake is detected?
An instant alert fires via email, Slack, or SMS with the confidence score and recommended action. Configurable automatic responses include meeting termination or additional authentication. A forensic report with frame-by-frame analysis is generated and available for export.
How does this work with GDPR and consent requirements?
Detection runs in real time against live audio and video. No call content is stored, which simplifies consent obligations significantly. The bot can be named and positioned as a note taker, which many organizations find falls within existing meeting recording consent policies. For deployments requiring explicit consent, notification preferences are fully configurable.
Can it detect candidates who use AI filters during interviews?
Yes. Face swaps, AI filters, and synthetic video personas are all covered by DETECT-3B Omni. Recruiters receive an in-call alert before the candidate advances to the next round.
What are the enterprise deployment options?
Cloud, on-premise, and air-gapped environments are all supported. Active Directory integration and SSO/SAML are available. SLA documentation available on request.
Get complete generative AI security
Join thousands of developers and enterprises securing with Resemble AI
