FREE AI SAFETY CHECKS / FOR PEOPLE

Before you trust a photo, an AI answer or an urgent voice call, check the evidence.

Q: Are these checks paid?

The reading guides and linked local tools are free. Where stated, they run in the browser and do not need a hosted model call.

A practical starting point for people who want a quick answer without creating an account, uploading a private file or sharing a call recording. Choose one question, read the short explanation and open the matching local tool.

photo privacyAI answer checkerEXIF GPSremove metadataAI image checkerurgent voice calls

Updated August 25, 2026FREE / PRIVACY-FIRST / OPEN METHOD

START WITH ONE QUESTIONThe fastest useful check is usually the one you can explain to someone else.

No score pretends to know the truth. Each path shows what the browser can observe, what it cannot prove and what to do next.

CHECK A SUSPICIOUS MESSAGE MAKE AN URGENT CALL PLAN CHECK A SITE BEFORE PAYINGFREE / NO ACCOUNT / NO HOSTED MODEL CALL

10-SECOND LOCAL CHECK

Try one sentence before you rely on the answer.

Paste a short claim. The browser only counts signals that deserve a source; it never decides whether the answer is true.

CHECK A SUSPICIOUS MESSAGE

AI checks for people / 7 STARTING POINTS

AI checks for people.

01 / PHOTO PRIVACYCan this photo reveal where I am?

Check GPS, device and other details before a photo leaves your phone.

CHECK PHOTO PRIVACY -> 02 / AI ANSWERSCan I trust this AI answer?

Separate evidence, assumptions and current claims before acting.

AUDIT AN AI ANSWER -> 03 / FILESWhat does this file reveal quietly?

Inspect EXIF and metadata locally, one field at a time.

VIEW METADATA -> 04 / SHARINGClean a photo before sending it.

Remove sensitive metadata and keep a copy ready to publish.

REMOVE METADATA -> 05 / IMAGESAre there signs of AI editing?

Review available provenance without turning uncertainty into a verdict.

CHECK AI IMAGE EVIDENCE ->06 / SUSPICIOUS MESSAGEA WhatsApp message asks for money or a code?

Sort the pressure and link signals, then copy a safe reply without sending the message anywhere.

CHECK THE MESSAGE ->07 / URGENT CALLSA familiar voice asks for money?

Make a short plan: end the call, use a saved number and do not send money or codes.

MAKE A CALL PLAN ->

Choose the risk you actually have

A photo can carry location data, an AI answer can hide an unsupported assumption, and a generated image can have provenance evidence that is easy to miss. Start with the object in front of you instead of a generic AI detector.

Keep private files local

The photo privacy, metadata and Content Credentials tools are designed for browser-side inspection. Read the page before selecting a file and do not treat a missing signal as proof that a file is safe or authentic.

Turn a result into a decision

Use the answer audit for dates, prices, current events and recommendations. Ask which claim is verified, which is inferred, what is missing and what observation would change the conclusion.

Share the explanation, not a fake certainty

A useful result names its evidence and its boundary. Save a clean copy only when you understand what was removed, and do not publish an AI probability as a verdict.

Use the next guide when the first result is unclear

The tools are connected as a reading path: inspect privacy, understand metadata, compare provenance, then audit the answer or image that caused the concern.

FAQ

Frequently asked questions

H/M Blindspot · hogarmas.net

Do these tools upload my photo?

The photo privacy, metadata and provenance workbenches are designed to analyze files in the browser. Read each tool's current privacy notice before using it; no local result should be treated as a server-side guarantee.

Can an AI detector prove an image is fake?

No. Available provenance or editing evidence can inform a review, but missing Content Credentials do not prove that an image is fake or human-made.

What should I check first?

If you are sharing a photo, start with privacy or metadata. If you are deciding whether to act on a response, start with the AI answer audit.

Are these checks paid?

The reading guides and the linked local tools are free. They run in your browser where stated and do not require a hosted model call.