HumanMeter for Fact-Checkers

Free, unlimited AI video and photo checks for fact-checkers and newsrooms. Paste a viral link, get a score and an explanation you can quote in seconds. We only ask that you name HumanMeter when it helps a published check.

Already in use on fact-check desks

HumanMeter has already been cited in published fact-checks, including Vishvas News' check of an AI-generated video falsely shown as floods in Nepal. In that August 2026 check, Vishvas News ran the video through Hive, HumanMeter and Sightengine, and published HumanMeter's reading (about 72% AI signal, with buildings changing shape and cars turning into water) as part of the evidence.

This program makes that official: unlimited checks for your desk, no paywall in the middle of a breaking story.

What your desk gets

  • Unlimited scans on humanmeter.app for every browser and phone you register.
  • Checks from a link: paste a public TikTok, Instagram, X or Snapchat link. No download needed.
  • Video and photo uploads up to 100 MB.
  • Quotable explanations: a verdict, an AI signal score from 0 to 100, and the top signal in plain English.
  • Provenance where it exists: content credentials, visible AI-generator watermarks such as Veo, and TikTok's own AI label are read and reported.

What we ask

When HumanMeter informs a published check, name it and link to humanmeter.app. That's all. No fee, no contract, no logo requirements.

Get access in two minutes

  1. Copy your browser ID below. Open this page on each browser and phone your team uses, and copy each one.
  2. Email it to us from your newsroom address, with a link to your published fact-checks.
  3. We switch it on within one business day. The badge above the scanner will say "Unlimited".
Your browser ID
loading…

Or write to support@humanmeter.app with your organization, a link to your fact-checks and your browser ID.

Use it as one signal, not the verdict

We'd rather you trust HumanMeter a little less and cite it accurately. It's strongest on fully AI-generated video and on media that still carries provenance. It's weakest on photos that were edited with AI and then re-saved or screenshotted, which is hard for every detector. Always pair a score with your own verification: the source account, the original upload, and reverse searches. See how HumanMeter works.

Questions

Yes. Verified fact-checkers and newsrooms get unlimited scans on humanmeter.app at no cost. There's no contract. We only ask that you name HumanMeter and link to it when a check informs a published piece.

Paste a public TikTok, Instagram, X or Snapchat link, or upload a video or photo file. You get an AI signal score from 0 to 100, a verdict (Likely Human, Unclear or Likely AI), a plain-English explanation and the top signal behind it. YouTube links aren't supported right now; download the clip and upload it instead.

Yes. The explanation is written to be quotable. Please describe it as a signal from an AI detection tool, alongside your own verification, not as proof on its own.

No detector is 100% accurate. HumanMeter is strongest on fully AI-generated video and on files that still carry provenance, such as content credentials, visible generator watermarks, or TikTok's own AI label. Images that were edited with AI and then re-saved or screenshotted are hard for every detector, including ours, so a low score means 'no strong signs found', not 'verified authentic'.

Uploaded video files are deleted when the analysis finishes. We keep the score, the link you scanned, and up to three low-resolution blurred frames so your history works. Nothing is published unless you share it. See the privacy policy for details.

Copy your browser ID from this page and email it to us from your newsroom address, with a link to your published fact-checks. We activate it within one business day. Each browser (and each phone) has its own ID, so send one for every device your team uses.