What a curated portal asks for before it accepts a game

There are two kinds of web-game portal. One accepts anything that passes an automated technical check. The other has a person look at every submission and says no often. This is about the second kind — what they check, in what order, and which item actually fails submissions.

Both kinds are worth having. The open ones are distribution and a little revenue from day one. The curated ones are a smaller number of much better placements, and being in one is a signal to the others.

The technical checklist, which everyone shares

None of this is documented in one place, but every curated portal checks the same things, and a submission that fails any of them does not reach a human:

• It loads fast. Under a few seconds on a mid-range phone. See the post on bundle size — this is the same constraint wearing a different hat.
• It works on a phone, in both orientations. Not "it runs" — it is playable, with controls a thumb can hit. A landscape-only game needs to say so and show a rotate prompt, not just break.
• It handles a resize. Rotate the device, and the canvas keeps up. A surprising number of games track the viewport once, on load, and never again.
• Audio starts on a gesture, never before. A game that plays sound on load fails browser autoplay policy anyway, but a portal treats it as a polish problem too.
• The tap targets are big enough. Roughly 44 px minimum for anything a finger has to press.

Get all of that right and you have cleared the bar that the open portals also use. You have not yet cleared the one that matters.

The originality bar, which is the one that fails you

This is the check nobody documents and the one that actually rejects games. A curated portal already carries fifty games in your genre. The question its reviewer is asking is not "is this good" but "is this another one of those". A competent, polished, completely standard entry in a crowded category is a soft no, and the rejection note will be vague because the real answer — "we have enough of these" — is awkward to write.

You can tell before you build. Describe the game in one sentence. If that sentence would fit twenty games already on the portal, the polish will not save it. Change the sentence, not the execution.

The assets that take longer than the game

• Covers at exact pixel sizes. Each portal wants its own set, and "close enough" gets bounced.
• A trailer whose first frame is not black. Video encoders love to open on a black frame; a portal's thumbnail grabber grabs frame zero.
• Screenshots with no debug overlay, no FPS counter, no dev console. Take them from a clean build, not your working one.

Budget real time for this. It is routinely the slowest part of a submission.

Read the exclusivity clause before you upload

Some portals ask for exclusivity, some ask for a licence to your IP, and at least one makes an upload effectively irrevocable. Read the terms for the words exclusive, perpetual, irrevocable and sublicense before the game goes up, because for some portals there is no taking it back down. Our own rule is simple: never exclusive, never a transfer of IP. It has cost us one portal so far and it was the right call.

The nine games we hold to that standard are on the originals page, and the terms we will and will not sign are on the licensing page.