For agencies and investigative teams

Media forensics for investigators

PersonaWatch Investigations gives verified agencies and investigative teams their own workspace to screen images, video, audio, headlines, emails and text messages for synthetic media and fraud, with recorded, hashed results, investigation reports designed with courtroom scrutiny in mind, and Fraud Alerts you can share.

Request access → How access works

Founded by two former FBI Special Agents.

What Investigations adds

Built for casework

Everything in PersonaWatch, plus a workspace and tools designed around how investigators work.

Investigations workspace

A separate workspace for case material. Every scan is recorded and stamped as an Investigations scan, with the SHA-256 of the file as received.

Investigation report

Case number, examiner and chain of custody, the finding and what it is based on, limitations, and a full methodology appendix. Screening indicators, not determinations.

Fraud Alerts

One-page alerts with handling markings (TLP or Law Enforcement Sensitive) to warn the public or partner agencies. Victim details are redacted on the server before export. You share them through your own channels.

Forwarded and uploaded email

Forward a suspicious email to PersonaWatch Investigations from your work email, or upload the email file (.eml or Outlook .msg). It appears in your workspace with a picture of the email as received, its links resolved and its claims checked.

Provenance checks

Content Credentials (C2PA) are read and validated, verified or self-declared. Provider watermarks (SynthID) are reported when present, including on screenshots where credentials were stripped.

Image History

Research an image's public circulation history: who published it, official attribution, fact-check coverage, and whether similar images circulate.

Six analysis engines

Fake, false and fraud

AI-generated images and video

S.A.R.A., PersonaWatch's proprietary detector, with generator attribution as an investigative lead and camera-metadata analysis.

Facial manipulation

Face-swap analysis when faces are present, with a face-region heatmap.

Voice clones

Synthetic and cloned speech detection for audio and the audio track of videos.

Headline fact-check

Claims checked independently by multiple AI models with web search, with their sources.

Phishing email

Sender, links and claims checked, with lookalike-domain and impersonation analysis.

Scam text messages

Text-message screenshots checked for scam patterns, link destinations and claimed organizations.

How it works

From evidence to alert

Analyze

Upload or forward the media. The engines run on the file exactly as received.

Record

Each result is stored as an Investigations record with the file's SHA-256.

Report or alert

Generate the investigation report, or a Fraud Alert with victim details redacted.

Share

Print or download, and share through your own channels. PersonaWatch does not host or distribute alerts.

How access works

Verified organizations only

There is no self-service sign-up. Access is set up for verified organizations, with seats for named members.

Request

Send the form below from your work email.

Verify

We verify your organization (for example, an agency contact or license).

Set up

We set up your organization's seats and contract dates. Pricing by agreement.

Sign in

Members sign in with their confirmed work email; the account, not a link, opens Investigations.

Request access

Tell us about your team

Work email only; personal email accounts (Gmail, Yahoo, Outlook.com and similar) can't be accepted. Please don't include Criminal Justice Information (CJI).