About · Est. 2023 · Montreal

We build the tools that tell you what's real.

Synthetic media is now good enough to fool anyone. We started Deepfake Detector because the people most exposed — families, journalists, finance teams — had no reliable way to ask a simple question: is this real, or is it a fake?

By the numbers

Built to give you a straight answer.

We've stayed deliberately lean and focused on one job: telling you whether a file is real. Here's what that focus delivers.

High
Accuracy
All
AI Models
Real
Human Detection
Free
Forever plan
Our story

It started with a phone call.

In 2023, a member of our founding team received a voicemail in the voice of a close relative asking for money. The voice was perfect. The relative had never made the call. That near-miss — and the realization that off-the-shelf tools couldn't have caught it — became the reason this company exists.

We set out to build detection that works the way a forensic lab works: not a single yes/no verdict, but evidence. Confidence scores, per-segment flagging, the specific signals that drove a call. Tools a journalist could cite, a fraud team could act on, and a court could understand.

Why we stay independent.

We don't build generation tools. We never will. Our only product is verification — which means our incentives point one direction: toward telling you the truth about a file, with no platform, model, or advertiser pulling the other way.

Today we run detection across image, audio, and video — for individuals, newsrooms, and finance teams — from our base in Montreal.

What we believe

Three principles.

They sound simple. They're the reason we turn down work that conflicts with them, and the standard we hold every release to.

01

Evidence, not verdicts.

We show our work. Confidence scores, flagged segments, and the signals behind every call — never a bare "fake" with nothing behind it. You make the call; we give you the basis.

02

Detection only.

We will never build synthetic-media generation. Our incentives stay pointed at the truth because we have no product that benefits from a fake going undetected.

03

Privacy by default.

Files are purged within 60 seconds of a verdict. We don't train on your media without explicit consent. What you submit is yours.

Milestones

The short version.

From a single detection script to coverage across every major generator. The road so far.

2023

Founded in Montreal

Started after a near-miss voice-clone scam. First audio detection model trained on a corpus of early voice-clone outputs.

2024

Image & video detection

Expanded beyond audio to cover AI-generated images and deepfake video. Launched the free Chrome extension.

2025

The API & enterprise

Shipped the public REST API and SDKs. Onboarded the first newsroom and finance-team customers.

2026

Forensic v3.2

Per-segment flagging, broader generator coverage, and a SOC 2 audit underway. High accuracy on our internal benchmark.

Detect Deepfakes Before They Spread.

Whether you're protecting a family, a newsroom, or a finance desk — start free. No card required.