How ChromaMimic keeps your frames on your machine, and how to check for yourself
"Nothing uploaded" is an easy thing to write on a landing page. This is what it means in ChromaMimic's case, in enough detail that you can check it rather than take our word.
Why it matters
If you shoot weddings, the frames are someone's private day. If you shoot commercial work, the frames are under NDA until launch. If you grade a feature, a still from the timeline on a third party server is a leak waiting to happen. Most LUT tools ask you to upload your footage because their processing runs on their servers, and their terms of service say what they can do with what you upload. Read them.
How ChromaMimic works instead
The site is a static bundle of HTML, CSS and JavaScript served from a CDN. When you drop a frame, the browser reads the file into memory from your disk. The colour analysis (Reinhard transfer, histogram matching, optimal transport) runs in a Web Worker, which is a background thread inside your browser tab. The preview is drawn by WebGL on your graphics card. The .cube is assembled as text in memory and handed to your browser's download function.
At no point in that chain is there a network request carrying image data. There is no upload endpoint to send it to. The only server side code in the product is the account and billing layer (sign in, subscription status, a download counter), and that layer never receives a frame.
The two minute test
- Open ChromaMimic and let it load.
- Open your browser's developer tools (F12, or Cmd Option I on a Mac) and switch to the Network tab. Tick "preserve log".
- Drop an original and a reference frame, press Build, and download the .cube.
- Look at the Network tab. You will see the fonts, the script chunks, and, if you are signed in, a small JSON call to the entitlements endpoint that carries your account state and nothing else. You will not see a request whose payload is an image.
Then the stronger version: load the page, turn your wifi off or set the Network tab to Offline, and repeat step 3. The build still runs and the download still works. A tool that uploads cannot do that.
What we do store
- If you sign in: your account (through Clerk), your subscription state (through Stripe) and a count of downloads this month.
- If you are on Pro and use the saved LUT library: the .cube files you choose to save, which are small text files, not your frames.
- Nothing else. No frames, no thumbnails, no histograms, no analytics that fingerprint your images.
The full detail is in the privacy policy, which is short enough to read.
The trade off
Running in your browser means the tool is bounded by your machine. A 65 point LUT with optimal transport on a very large frame takes a few seconds on a laptop and longer on an old phone. Server side tools can throw more hardware at it. We think a few seconds is a fair price for never handing a frame to anyone, and the progress bar is there so the tab never appears frozen.
Questions
- Does ChromaMimic use my images to train anything?
- No. It never receives them. The colour matching is classical colour science, not a model that learns from users.
- Can I use it on a machine with no internet at all?
- Once the page has loaded, yes. It needs a connection to load the first time and to check a signed in account's download allowance.
- Where is the account data hosted?
- Authentication is handled by Clerk and billing by Stripe. ChromaMimic's own server functions run on Vercel and hold only your entitlement state.