Back to audit
Plain-English commitments

How we handle your video.

You're trusting us with unreleased work. We don't take that lightly. Here is exactly what happens to your video — in plain language, backed by our code.

What we receive

When you upload a video, the file streams directly to our forensic backend over an encrypted HTTPS connection. It is held in a temporary, isolated sandbox for the duration of the audit — nothing more.

When we delete it

The moment your audit finishes — whether it succeeded, failed, or you cancelled — we delete the file. Deletion runs in a finally block, so it happens even if something errors out. Typical retention: 1–3 minutes.

What we never do

We never store your video on permanent disk. We never share it with third parties. We never use it — or any audit response — to train any model, ours or anyone else's.

What we keep

Only the bare metadata required to operate the service: anonymous rate-limit counters, structural error logs without file content, and aggregate usage stats. No filenames, no thumbnails, no transcripts.

Technical details

How analysis actually works

  1. 1. Upload. Your browser sends the file to our backend via TLS. We stream it to a temporary path on disk (random filename, isolated working directory).
  2. 2. Analysis. The file is sent to a forensic analysis engine that scores the content for inauthentic-content signals.
  3. 3. Deletion. The temporary file on our backend is removed immediately.
  4. 4. Display. You see the score and report in your browser. Close the tab and it's gone.
The bottom line

Your unreleased video stays yours. We are a temporary forensic checkpoint — not a content library, not a training dataset, not a leak vector. If our policy ever changes, you'll see it announced on this page before it takes effect.

Last updated: June 7, 2026.