Valve changes these rules. Treat this as the shape of the process, then confirm the specifics on Steamworks before you commit money or a launch date.
Having a finished build is roughly half of publishing on Steam. The other half is paperwork, waiting, and a few rules that quietly decide whether your game makes money. Here's the whole path in order.
Step 1. The Steam Direct fee
Steam charges a $100 fee per game. It's recoupable: once the game earns a certain amount in adjusted gross revenue, Valve credits it back. If your game never gets there, the $100 is gone.
Two consequences people underestimate. First, this is per title, not per account. Five games is five fees. Second, it's a real filter: a tool that makes games cheap to produce does not make them cheap to publish.
Step 2. Identity and tax paperwork
Before you can take money, Valve needs to know who you are and how to pay you. That means identity verification, bank details, and a tax interview: a W-9 if you're in the US, a W-8BEN or W-8BEN-E if you're not.
If you're outside the US, this step decides how much withholding comes off your revenue. Whether your country has a tax treaty with the US, and whether you filled the form correctly, can be the difference between keeping most of your earnings and losing a chunk at source. It's worth thirty minutes of care.
Step 3. The waiting periods
This is where launch plans die. Steam enforces waiting periods that you cannot shorten:
- After paying the fee, there's a mandatory wait before you're allowed to release anything.
- Your store page must be live for a minimum stretch before launch day. Valve wants a "Coming Soon" page collecting wishlists, not a game appearing out of nowhere.
- Every build you submit goes through a review by Valve that takes business days, not minutes.
Add them up and you should be starting the Steam side weeks before the date you want to launch, not the week of. If you take one thing from this post: create the account and pay the fee early, even if the game isn't finished.
Step 4. The store page
Valve has specific requirements, and a page that's missing pieces won't get approved:
| Asset | Why it matters |
|---|---|
| Capsule images (several sizes) | This is your ad everywhere on Steam. It does more for sales than the trailer. |
| Screenshots | Must be actual gameplay, not marketing art. |
| Trailer | Gameplay in the first five seconds. No long logo intro. |
| Short and long description | The short one is what people read. Write it last, and write it plainly. |
| System requirements | Honest ones. Overstating causes refunds. |
| Content survey / age rating | Violence, language, and so on. Answer accurately. |
Step 5. The AI disclosure
Since 2024, Valve requires developers to disclose their use of AI in the content questionnaire: what was generated, and whether it happens pre-baked or live while the game runs. That disclosure is shown on your store page.
Two honest observations. Practically, disclosing is straightforward, and games that disclose get published. Culturally, some players do react to the label, and there is no point pretending otherwise. The games that get a hard time are usually the ones where the AI-generated content is the whole product; the ones that don't are the ones where it's a tool in service of something considered. Answer the questionnaire truthfully. A false answer is a much worse problem than a label.
Step 6. The refund rule that catches small games
Steam grants automatic refunds for purchases under roughly two hours of playtime made within the last two weeks. No reason needed.
Read that again if you're making a short game, because it's the single most expensive fact in this post. A polished 90-minute game is a game a player can finish and then get their money back for, entirely within the rules.
Realistic ways people handle it:
- Price short games low enough that refunding isn't worth the friction.
- Build in replay value (modes, scores, difficulty tiers) so two hours isn't the whole thing.
- Be explicit about length on the store page. Players who know what they bought refund less.
- Consider whether Steam is even the right store for a 45-minute game. itch.io and mobile have different economics.
Step 7. Launch, and the part that isn't Valve's fault
Uploading the build is the easy part. The uncomfortable truth is that Steam is a marketplace with more games released every day than anyone can look at, and visibility is mostly earned before launch, not after, largely through wishlists accumulated while the Coming Soon page sat there.
A game that launches with a handful of wishlists usually stays where it launched. That's not a punishment for being small; it's just how the algorithm reads early signal. Which is one more reason to get the store page up early and point everyone you know at it.
The short version
| What | Reality |
|---|---|
| Cost to publish | $100 per game, recoupable |
| Whose account | Yours. Never publish under someone else's. |
| Time before launch to start | Weeks, not days |
| Build review | Business days, every submission |
| AI content | Must be disclosed, and it's shown publicly |
| Biggest hidden risk | Refunds on games under ~2 hours |
| Biggest real risk | Launching with no wishlists |
None of this is a reason not to publish. It's a reason to start the boring parts early: the fee, the tax form and the Coming Soon page all wait better than you do.
Fliper Studio produces the build and the store files and walks you through this list, but the account, the fee and the review stay yours, on purpose. Here's why that matters.