How Photographers Can Track Client Proofs and Previews

Short answer: Proofs and previews often leave your hands before the client has paid or signed off. A visible watermark protects them from casual reuse, but it looks the same on every copy, so it can't tell you who shared a photo. If you also give each client a copy with its own invisible watermark, a leaked photo can be checked and matched to the copy it came from.

Last updated: October 9, 2026.

Where proofs and previews leak

A typical job involves several people who see unfinished work: the client, their team, a second shooter, an agency, sometimes a venue or a model. Photos can leak at any of these points:

  • Online galleries where someone saves the images or takes a screenshot.
  • Email or messaging threads forwarded to people you never met.
  • Pre-delivery previews posted early to social media by a client who was just excited.
  • Embargoed or commercial shoots where early images matter to a launch or a campaign.

Most of the time it is not malicious. But when it matters, such as an unpaid job, an exclusive shoot or an embargo, you want to know where the file came from.

What a visible watermark can and can't do

A visible watermark, such as your logo or a diagonal text across the image, is still a good idea for public previews. It discourages reuse and gives credit. Its limits are well known:

  • It can be cropped, cloned out or painted over.
  • It spoils the picture for the client, who is trying to judge your work.
  • It is identical on every copy, so it says "this is mine" but not "this client shared it."

For a longer comparison, read visible vs invisible watermarks.

The approach: a different invisible mark per client

An invisible forensic watermark is a pattern written into the pixels that you don't see at normal viewing size. If each client, or each recipient, gets a copy with its own pattern, then every photo in that set carries a label that points to that copy, even after it has been renamed. A re-upload, a resize, or a screenshot can still match when enough of the mark survives.

This does not stop anyone from sharing. It gives you evidence when something leaks, and when clients know their copies are traceable, they tend to be more careful.

How to set it up

  1. Choose what to mark. For paid previews and proofs, mark the images clients actually receive. Marking the whole set is better than marking a few.
  2. Create one marked copy per client or recipient. With Markedfile you upload an image and create a copy for each recipient, labeled with something neutral like a job number, not a name.
  3. Deliver each client their own set. Use your usual gallery, Drive folder or email link, but make sure each link carries that client's copy.
  4. Keep a record. Note which label went to which job and recipient. Keep it outside the file.
  5. If a photo leaks, check it. Download the leaked image, upload it to Detect, and see which copy it matches.

Files are processed in memory and are not stored. The Free plan covers 5 active files with detection for 6 months, and paid plans start at $9.99/month.

What to do after a match

A match tells you which copy a photo came from. It doesn't prove who posted it: a client may have shared a gallery link, a colleague may have taken a screenshot, or an account may have been compromised. Use it as a starting point.

  • Ask the client first, politely and with the details of where you found the photo.
  • Update your contract or gallery terms to say that proofs are individually marked and that sharing before delivery or payment isn't allowed. Telling clients in advance is what makes the mark useful as a deterrent.
  • Pause access to the gallery or the final delivery if your terms allow it.
  • Request a takedown of the leaked image if you own the rights and it is hosted publicly.

Where this approach has limits

  • Detection is never guaranteed. Strong editing, heavy compression and very small images can destroy a mark. Small images and some heavily compressed ones use a more fragile mark that does not survive screenshots; images under 256×256 px always do. A crop of roughly 230 px of a detailed area is the approximate minimum, not a ceiling. Images marked before the latest resize improvement need to be marked again before a resized copy can resolve, and a resize can still miss on a heavily textured image.
  • Only marked files can be traced. Photos you sent unmarked in the past can't be matched later.
  • It identifies a copy, not a person. Be careful before accusing anyone.
  • Clients can ask for the clean file. That is fine: mark the final delivery too, with its own label.

FAQ

Will the watermark change how my photos look? It is designed to be invisible at normal viewing size. Compare the original and the marked copy at 100% on a screen you trust before relying on it for client work.

Can a screenshot of my proof still be matched? In many cases, yes, since the mark lives in the pixels and part of it can survive. It depends on how much of the image is left and how detailed the area is, and detection isn't guaranteed.

Should I still put a visible watermark on public previews? Yes, for anything posted publicly. Use the invisible mark for what you send privately to clients.

Can I use a client's name as the label? It's better not to. Use a job number or order ID and keep the mapping in your own records.

What about RAW files? Markedfile works with image files and PDFs. Mark the exported JPEG or PNG you send to clients.

Does it work for weddings and events with hundreds of photos? You can mark a large set, but check the limits of your plan: Free covers 5 active files, and paid plans raise that.

Try Markedfile free · For teams

Mark each copy before you share it

Create a free account. You can mark 5 files and see whose copy leaked.