Skip to content
TraceRealAI

Detect what enters your organisation.Protect what leaves it.

A rotating set of media receives two independent TraceRealAI checks. Each image is identified as manipulated, with a concise description of the edit, and checked for an invisible TraceRealAI watermark.
7+ yearsMedia integrity research
20+ papersCVPR, NeurIPS, ICML & ACM MM
MultipleBest Paper Awards
World’s largestProprietary deepfake dataset owners

How TraceRealAI works.

Detect manipulation and check watermarks.

Upload images, video or audio. TraceRealAI checks for manipulation and, for protected images and video, a TraceRealAI watermark. Open the result to see the affected region, timestamp or supporting explanation.

  1. 01Upload
  2. 02Analyse
  3. 03Review result
Incoming media
A conference recording frame with an altered pointing gesture highlightedVIDEO · 00:18
00:00
00:36
Detection resultManipulation detected

A speaker’s gesture was changed to point directly at the viewer.

conference-recording.mp4
Switch between a video edit and an image edit. Each example shows the media region and plain-language explanation behind a manipulation result.

Add an invisible watermark before publishing.

Add a TraceRealAI watermark to images and video without placing a visible badge over the media. Publish as normal, then check a later copy to see whether the marker is still present.

  1. 01Add watermark
  2. 02Publish
  3. 03Verify later
Outgoing media
Watermark appliedMarker verified
A comparison slider moves across visually identical original and watermarked media, followed by concise confirmation that the invisible marker was applied and verified.

Manipulated media can trigger real decisions.

Images, video and audio are used to approve payments, confirm identities and represent organisations. A convincing file can still be altered or impersonate someone your team trusts.

8Mestimated deepfakes shared in 2025

Deceptive media is scaling faster than manual checks.

That is sixteen times the estimate from two years earlier. The UK Government says the figure spans false images, video and audio.

Global estimate · cited February 2026UK Government case study
HK$200Mreported transferred after a fraudulent video conference

A familiar face on a video call can still be fabricated.

In Hong Kong, an employee joined a prerecorded video conference that impersonated the company’s chief financial officer and colleagues, then transferred the funds requested during the call.

Hong Kong case · reported June 2024Hong Kong Government case report
43%had encountered at least one deepfake online in the previous six months
9%felt confident identifying one

Visual inspection is not a dependable safeguard.

In Ofcom’s study, 43% of UK adults said they had encountered a deepfake in the previous six months, but only 9% felt confident they could identify one.

UK adults aged 16+ · June 2024 · n=2,000Ofcom research
May 2026Singapore Police released footage from a deepfake Zoom scam

Government leaders were impersonated in a fabricated meeting.

Singapore Police released footage of a call that combined recorded video, fabricated images and inauthentic audio. A supposed lawyer later requested a transfer.

Singapore · police case reportSingapore Police case report
22,364complaints that referenced AI
US$893Min associated reported losses

Fraud is no longer limited to one format.

The FBI recorded 22,364 complaints that referenced AI in 2025, with US$893 million in reported losses. The category includes voice clones, fake profiles, documents and fabricated video.

United States · 2025 · broad AI-related complaint categoryFBI Internet Crime Report release

TraceRealAI fits where your media already moves.

Use the same detection and watermarking capabilities through a web workspace, automated APIs and integrations across review and publishing tools.

Images, video and audio flow through TraceRealAI for detection, explanation, watermarking and verification. The capabilities are available through a web workspace, APIs and plugins.
01Web workspace

Check or protect media directly when someone on your team needs an answer.

02APIs

Automate detection, watermarking and verification inside products and media pipelines.

03Plugins and integrations

Bring TraceRealAI into the review, publishing and collaboration tools your teams already use.

Try TraceRealAI with your own media.

Open demo