Academic Pricing: Why Your Email Is Not the Proof

Your email address does not get you the discount here, and that is deliberate. Academic pricing is granted on a document, a current transcript or a student or staff ID card. Your institutional address is still required, just not as the proof: it is checked against the document, and an application arriving from a consumer mailbox is never approved automatically. Corroboration is what an address is actually worth, and it is worth that only when it corroborates something.

Why the address test fails twice

The obvious implementation of student pricing is a pattern match on the address: does it end in .edu, or contain .ac. That works well in two countries and fails everywhere else. Checking our own signups, real universities issue mail on ordinary national domains constantly. Universidad de Buenos Aires uses fi.uba.ar. Universiti Teknologi Malaysia uses graduate.utm.my. Neither matches any pattern short of a list, and a list is never finished.

That is the first failure, and it lands hardest on people whose institution is already less legible to software written elsewhere. The second failure runs the other way: disposable .edu addresses are sold openly, for a few dollars. So the same test that turns away a real student in Buenos Aires waves through anyone who bought an address this morning. A check that is both too strict and too loose is not a check. A document shows a name, an institution and a date, and it is what we ask for.

What the document review actually does

An automated first pass reads the upload immediately. It is worth being precise about what that pass is and is not, because the honest version is less impressive than the marketing version.

It is not a forgery detector. A small vision model cannot tell a genuine student card from a good fake, and asking it to try produces a confident answer with nothing behind it. What it does is read what is legible: is there an institution name, does it look like a card, is there a date. A policy then decides, and every time the policy overrules the model, that is recorded.

It can approve, and it can escalate. It rejects only a few unambiguous cases. The asymmetry is deliberate. A wrong approval costs a few dollars a month. A wrong rejection costs a person their access, and falls hardest on a real card in a script the model reads poorly, photographed on a cheap phone. When it is not sure, a human looks.

What happens to the document

An ID is the most sensitive thing anyone uploads here, so: it goes to a private location that the public asset CDN cannot read, it is streamed to a reviewer through the API rather than handed out as a link anyone with the URL could open, and it is deleted a day after the decision. The verdict is kept. The evidence is not.

What approval gets you

A price, not a plan. This is worth saying plainly because it is the part people misread: being approved does not upgrade your account, it makes the academic rate available to buy. Everything free stays free either way. Unlimited collaborators are not a paid feature here and never have been, which is the difference that matters most to a lab, since the usual model charges per seat and turns a six-author paper into six subscriptions.

Whole departments

If you are asking on behalf of a department rather than yourself, individual verification is the wrong mechanism. A site licence covers everyone with an address at the institution, with no per-person approval step at all. Students and researchers at 240 or so universities and research labs across more than fifty countries already write here, so there is a reasonable chance yours is among them. Start at the academic page, and choosing an editor for a university covers what to ask before standardising a department on anything.