What happens to your file

Your file is read by JavaScript and WebAssembly running in your own browser. It is never sent to us. There is no upload request to intercept, no bucket it lands in, and nothing to delete afterwards. Closing the tab is the deletion. You can check this yourself: open the network panel, convert a file, and watch that no request carries it.

Rebuilding a solid from a mesh needs more memory and time than a browser tab can reliably give it, so it runs on a server. The page says so before you choose a file. The upload is stored in object storage under a lifecycle rule that deletes it after one hour, and the result is deleted with it.

A job row exists so the queue can run and so we can measure what conversions cost. It holds the format converted from and to, the size in bytes, the triangle count, a status, an error code if it failed, and how much processor time it used. It holds no IP address, no filename, no account and no file contents. There are no columns for them, so there is nothing to forget to delete.

One more table exists for this lane: a rate limit of ten uploads per address per hour. It stores a salted hash of the address and the hour, not the address, links to no job, and is pruned within two hours. It is the reason the upload path cannot be used to run someone else's server bill up.

Every capability claim on a format page (what a format stores, which applications open it) carries a link to the vendor's or standard's own documentation and the date it was checked. Where a source doesn't confirm something, the page says "unconfirmed" and keeps the value visible. Where sources disagree, it says "disputed" and shows both. An inferred yes is never published.

Every conversion claim is generated from a test run rather than written by hand. A pair that fails its tests has no page.

After a conversion or a viewer open, the page reports what happened in counts only: the two formats, the file's size in bytes, its triangle count, how long it took, and whether it worked. That is what lets us see which conversions people use and how large the files are. It does not include the filename, the file, your address, your browser, or anything that would tie two reports to the same person, and the endpoint that receives it reads no headers. The aggregates are public at/api/stats.

Google Analytics measures page views and how visitors use the site. It may use browser identifiers such as cookies; it never receives the CAD files you open or convert in your browser. There are no accounts: nothing to sign up for and nothing to sign in to.