Steam supported languages: what interface, audio and subtitles mean

At some point in setting up a Steam page you reach a grid: every language Steam supports down one side, three columns across the top, and a lot of empty checkboxes. Nothing on the screen tells you what standard you are being held to. There is no minimum coverage percentage, no review step, no automated check. You tick what you tick, and the store shows it to everyone who looks at your page.

That absence of a rule is why the question comes up constantly. A game with translated menus but original-language story text: does that count? A translation done with a machine and skimmed once: does that count? A language a fan contributed and you cannot read: does that count?

This article works through what each of the three columns actually claims, where a reasonable line sits for calling a language supported, and what happens on both sides of that line, because over-claiming and under-claiming both cost you something and the costs are not symmetrical.

What the three columns actually mean

The three columns are Interface, Full Audio, and Subtitles, and they describe three different things about your game rather than three grades of the same thing.

Interface means the text a player reads to operate the game: menus, settings, HUD, tutorials, item and ability names, error messages. In practice players read this box as a promise about all the text they will encounter, which is where the most common mismatch comes from. If your menus are translated but your dialogue is not, ticking Interface tells a player something you have not delivered, even though it is arguably accurate on a narrow reading.

Full Audio means recorded voice in that language. The word doing the work is full. Voice in cutscenes only, or a few barks, is not full audio, and claiming it produces a specific and very visible kind of disappointment when the first unvoiced scene arrives.

Subtitles means subtitles for spoken content in that language. It only means something if your game has audio to subtitle. A game with no voice acting has nothing to claim in either audio column, and players reading a row with only Interface ticked will generally understand that correctly, provided the Interface claim is true for all the text.

The columns are independent. A game with English voice acting and a full Japanese text translation ticks Interface and Subtitles for Japanese and Full Audio only for English. That combination is normal and reads clearly to players.

Nobody verifies this, which makes it a promise rather than a setting

It is worth being blunt about the mechanics: no one at the platform plays your game in every language you have ticked. The checkbox is not a certification. It is a statement you are making to buyers, and it is enforced the way statements to buyers are always enforced, which is by the buyers.

The enforcement mechanism is refunds and reviews. A player who buys because the store said their language was supported, finds it is not, and refunds costs you nothing directly. A player who does the same and writes a review costs you something that lasts, because reviews stay visible long after you have patched the problem and almost nobody goes back to revise one.

This is also why store filters matter in the other direction. Players can browse and filter by language, and many keep their store preferences set so that games without their language are less prominent or hidden entirely. An unticked box can remove you from consideration before a human ever sees your capsule art. So the box is doing real work in both directions, which is exactly why the decision deserves more than a shrug.

Where the line is for calling a language supported

The useful test is not whether the translation is excellent. It is whether a player who reads only this language can play your game from the first screen to the end without switching languages and without hitting text that is missing, unreadable, or in a language they do not know.

That resolves into three floors that are worth checking explicitly before you tick anything:

• Coverage — no untranslated strings anywhere on the main path, including the parts nobody tests: the settings menu, the pause screen, save and delete confirmations, network and error messages, the credits, and the text on the store and DLC prompts
• Correctness where it gates progress — control prompts, tutorials, puzzle hints, anything that tells a player what to do. A charming mistranslation in flavour text is a shrug; a wrong control prompt is a refund
• Rendering — no missing glyphs, no boxes where characters should be, no text running past its container or clipped by a button. A perfect translation that does not fit its UI is not a supported language

The cost of ticking too many boxes, and of ticking too few

Over-claiming fails in a way that is disproportionate to the size of the market you were reaching for. The players who arrive because of the checkbox are precisely the players who cannot tolerate the gap, and they arrive already expecting the thing you did not deliver. Their reviews are written in that language, which means the next player from that market reads a warning in their own language before they read your description. Sales from a language you claimed badly can be net negative once the review score damage is counted.

Under-claiming fails quietly, which is what makes it hard to notice and easy to over-correct for. Players who would have bought never see the page. You get repeated questions in the discussions asking whether the game supports a language it partly does. Nothing dramatic happens; you simply do not get the sales.

The asymmetry is the whole decision. Adding a language to the table later is trivial: you patch, you tick the box, you post an announcement, and the announcement itself is a small marketing moment. Recovering from a wave of reviews saying the translation is broken is not trivial and may not be possible. So when you are unsure, the direction to err in is fewer languages done properly, and the direction to move over time is to add.

How to present partial support honestly

Partial support is extremely common and it is not something to hide. It is something to state, in the language it applies to, where the person it applies to will read it.

The important detail is the language of the disclosure. A note in your source language explaining that the Japanese translation covers menus only does not reach the Japanese-reading player who needs it, because that player is not reading your source language. Put the note in the language it concerns, near the top of the description on that language's version of the page. Steam also gives you a place to attach a short note alongside your language settings; check the current Steamworks documentation for exactly where, and use both.

Disclose the method too when it is unusual. Players are markedly more tolerant of a machine-assisted or community-contributed translation that is labelled as one than of the same translation presented as professional work. The label converts a disappointment into an expectation, and it usually converts complaints into corrections, because people who know a translation is community-made tend to offer fixes rather than write reviews.

Language support (as of the latest update)

English    - full: interface, all story and item text
German     - full: interface, all story and item text
Japanese   - interface and menus only; story text is in English
Portuguese - community translation, in progress; corrections welcome at <link>

Full voice acting is English only. Subtitles are available in every
language listed above.

Adding a language: the order of operations

The box is the last step, not the first. Ticking it before the rest is done is how the bad version of this goes, because the store starts sending you players immediately while the build is still catching up.

• Confirm the text is complete — run a check for missing or untranslated keys rather than trusting that the file looked finished
• Play the language end to end, on the platform you actually ship on, at your minimum supported resolution and window size
• Fix rendering problems found in that pass: font coverage, overflow, line breaking, anything clipped
• Update the store page in that language, including the short description
• Ship the build, then tick the boxes and post an announcement in that language
• Watch the reviews and discussions in that language for the first couple of weeks, because that is when you find out what your own testing missed

One decision to make before any of this

Everything above assumes you have already chosen which languages to attempt. That choice deserves its own thinking, and the checkbox grid is a bad place to make it, because a grid of empty boxes invites you to fill them.

The healthier frame is that each row is a commitment with a tail: every language you tick is a language every future patch, event, and piece of DLC has to be translated into, forever, or the row quietly becomes false. A language you cannot keep up with is worse than one you never started, because the game degrades in that language over time while the store keeps promising otherwise. Count the ongoing cost before you count the market.