Built to defend trust in live conversations.

Diopter protects teams from AI social engineering on video and voice calls, so the person answering does not have to be the last line of defense.

Mission

AI is not the threat. Deception is.

Synthetic video and cloned audio make it harder to trust the calls where teams approve payments, grant access, verify claims, and make hiring decisions. Diopter exists to make those moments transparent, actionable, and defensible.

Founders

Founded by builders who understand both AI and abuse.

Diopter was founded by Rohan Gupta and Jeremy Pippin after seeing how quickly generative AI moved from productivity use cases to impersonation, fraud, and social engineering. Their focus is simple: make defense as accessible as the attack.

RG

Rohan Gupta

Co-founder & CEO

Founder and former CEO of QuillBot, a generative AI company with 35M+ monthly users.

JP

Jeremy Pippin

Co-founder & CPO

Former VP of Product at QuillBot and FanDuel.

What we believe

Four things shape how we build.

01

Humans have a right to know.

People should not have to guess whether a voice, face, or urgent request on a call is real.

02

The arc matters.

Fraud is rarely detectable from a single frame or voice sample. It unfolds across a conversation, and the arc is the unit of analysis.

03

Verification should feel empowering.

Diopter should help the person on the call feel protected, not watched, judged, or restricted.

04

Defense should be as accessible as the attack.

AI impersonation is becoming cheap and scalable. Protection has to be deployable for the teams attackers actually target.

Scope

Where Diopter focuses today.

A clear scope for evaluators. Diopter focuses on live video and voice calls and the conversation arc that runs through them.

Video calls

Synthetic video, identity drift, proxy interviewers, and executive impersonation across conferencing.

Voice calls

Cloned audio, vishing, caller impersonation, and high-pressure approval calls across phone and VoIP.

Conversation arc

Authority, urgency, isolation, escalation, and the ask, scored continuously across live calls.

Deployment

Endpoint and voice-path coverage, rolled out through the MDM your IT team already runs.

Security & privacy posture

Built for sensitive environments.

Diopter is designed to be deployed inside the perimeter of teams that can't ship raw call media to a vendor cloud. Detection runs on the endpoint by default; only verdicts and scores leave the device.

  • SOC 2 Type II in progress
  • Configurable data retention
  • Tenant-level controls
  • Granular admin policy
  • No caller-side install where applicable
  • Data minimization by default
Security contact security@diopter.ai
Who we serve

Enterprise-grade defense, built for every team attackers target.

Diopter is built and tested to the bar enterprise security teams expect, and packaged so mid-market and SMB IT & security teams can deploy it through their existing MDM in days, not quarters.

Enterprise security & fraud

Built to enterprise standards. SOC 2 Type II in progress, tenant-level controls, MDM-ready rollout, and integrations into the SOC and case workflows you already run.

Mid-market fraud & security

The same detection engine, deployable in days through your existing MDM. Built for fraud, security, and IT teams owning the calls where money and access move.

SMB IT & security teams

Lean teams that need enterprise-grade defense without a dedicated SOC.

Contact

Talk to Diopter.

Walk through a real deepfake incident and see how Diopter scores identity, synthetic media, and the conversation arc.

Walkthrough · 30 min · NDA-safe

Walk an attack arc with Diopter.

In 30 minutes, we will replay a real deepfake incident, show the signals Diopter would score, and map the verdict your team could act on.