If you make art, you've felt it: the quiet dread that the moment you post your work, it gets vacuumed into an AI training set and spit back out in a thousand cheap imitations of your style. The instinct is to ask for a wall — something that stops it. The honest truth is that no tool stops it completely. But two tools, used together, change the game.
The trick is to stop looking for one magic wall and start thinking about two different jobs: making your work harder to copy, and proving it was yours when someone copies it anyway. Those are different problems. They need different tools.
Two problems, two tools
Problem 1 — mimicry. AI models learn your style by studying your pixels. Cloaking tools fight this by adding subtle, mostly-invisible perturbations to your image — changes a human barely notices, but that confuse a model trying to learn your style.
Problem 2 — priority. Even a perfectly cloaked piece can still be screenshotted, reposted, and claimed by someone else. Cloaking does nothing for that. For that you need proof: a permanent, public, timestamped record that you made this, first.
So the workflow that actually protects artists is two steps: cloak it, then prove it.
Step 1 — Cloak it (Glaze)
The leading free, independent cloaking tool is Glaze, built by researchers at the University of Chicago specifically to defend artists against AI style-mimicry. It runs on your own computer, adds its protective layer to your image, and you post the cloaked version instead of the raw one.
You can download it from the team's official site: glaze.cs.uchicago.edu. (We're not affiliated with Glaze or the University of Chicago — we just think it's the best free tool for this job and we'd rather point you to it than sell you a worse version.)
One honest caveat, because we're an honesty company: cloaking is best-effort, not a guarantee. It raises the cost and difficulty of mimicry — it does not make your work impossible to copy or learn from, and researchers have already shown that some cloaks can be partially reversed as models improve. That's not a reason to skip it. It's a reason not to stop at it.
Why cloaking alone isn't enough
Cloaking is a moving target — an arms race between protection tools and the models they're trying to fool. Today's cloak may be weaker against next year's model. And cloaking is silent: if it fails, nothing tells you, and you're left with no defense and no evidence.
That's the gap. Cloaking tries to prevent the theft. It can't give you anything to hold up after one happens. That second job — the receipt — is where a seal comes in.
Step 2 — Prove it (OriginProof)
Sealing your work with OriginProof creates a permanent, public record that this exact piece existed at this exact moment, tied to you. It takes about a minute and you get 10 free seals a month — 20 your first month. When you seal, you get an independent verification link anyone can check — no account required to look.
That proof is the thing cloaking can't give you. If someone reposts your art and claims it, you don't argue — you paste the verify link that shows your seal predates their post. If an AI detector wrongly flags your human work as machine-made, your timestamp shows the piece existed before the model that supposedly made it. The proof is anchored so it can't be quietly backdated or altered, and it stays verifiable for good.
Cloaking is the lock on the door. The seal is the receipt in your pocket. You want both — but if you only keep one, keep the one that still works after the lock is picked.
The habit: Cloak it. Prove it.
- Cloak it. Run your finished piece through Glaze and save the protected version.
- Prove it. Seal that version on OriginProof before you post it — so your public proof matches exactly what the world sees.
- Post it. Now your style is harder to copy and you hold a timestamped receipt if anyone tries to claim it.
It adds maybe two minutes to your process. In exchange, you stop being the artist with no defense and no proof — and become the one with both.