Patchworld is in-game localization QA, by the hour. Native players play your real build on real devices, in their own language and culture, and hand you a clean bug list. Send a TestFlight, Google Play or Steam link. Buy testing hours. Patch what breaks.
From $30/hour · 1-language minimum, buy from 1 hour · No sales call · 100% human native testers
In August 2025 Steam made per-language review scores the default, and called out "translation issues" and "cultural references" by name. Your localization quality is now a public, per-region number that players see before they buy. You cannot hide a broken language behind a good one anymore.
of Steam monthly active users are not English-first. Most of your players read your game in a language you do not speak.
positive: where Silksong's Simplified Chinese reviews landed after a tone-deaf translation, on ~35,000 reviews.
text expansion for German vs English. The single most common loc bug is a button that no longer fits its label.
of your translation spend is what LQA costs. It is the part that keeps the other 80-90% from looking amateur.
You already know how to ship a beta. We plug into the channel you already use, our native testers play it, and you get bugs back the way your devs like them: in your tracker, with screenshots and repro steps.
TestFlight, Google Play internal testing, a Steam playtest key or an APK. We get onto your real build, on real devices.
Real native-speaking gamers play your game in context: every menu, quest and edge case. We send your glossary and style guide along, not just a string table.
Every issue tagged by type and severity, with a screenshot, the in-game location, repro steps and a proposed fix. Straight into Jira, Trello, a sheet, whatever you use.
You fix, we run regression on a fresh build to confirm each fix actually landed. New content next month? Send the patch, we test the delta. No re-onboarding.
A translation can be perfect in the string table and broken in the game. These only show up when someone actually plays the localized build, in the target language, on a real device.
Text that outgrows its box. German, Finnish and Russian expand; buttons clip, tooltips cut off, numbers overlap labels.
Missing glyphs render as empty boxes; wrong encoding garbles symbols. Your font almost certainly lacks CJK, Cyrillic, Arabic or Thai.
Arabic and Hebrew need shaping and a mirrored UI. Letters that do not join, bars that fill the wrong way, back buttons on the wrong side.
Dynamic strings that never resolve, or grammar that breaks when the engine cannot move the variable for the target language.
English has 2 plural forms; Russian 3, Polish 4, Arabic 6. Gendered languages change verbs and adjectives. Naive "{n} items" is wrong almost everywhere.
Symbols, religion, maps and taboos that read as a ban or a recall, not a typo. A skull in a China build, a verse under a fight scene, a flag on the wrong border.
Culturalization goes beyond translation. It checks your game's symbols, history, religion and geopolitics against what each market will actually accept. These failures only reveal themselves when someone sees the scene, never in a string list.
Our native testers flag them while you can still fix them in a patch, not after the hashtag trends.
Almost every LQA vendor hides behind "request a quote". We do not. One clean hourly rate, a 1-hour minimum, and you can estimate the cost yourself before you ever talk to us.
A typical indie title runs ~8-15 hours per language; mid-size ~20-40. Drag to get a ballpark.
A string checker reads your text out of context. A player meets it in a box, at a moment, on a screen. Only one of them finds the bug your actual players would.
CJK line-breaking, RTL shaping, Cyrillic and diacritics: the scripts that break naive code are exactly the ones we test hardest, with native speakers who actually game in that language.
languages and locales covered
human native testers, never raw MT
platforms: iOS, Android, Steam, console builds
per hour, public, no sales call
Send us your build link and the languages you care about. We'll come back with a plan and an hour estimate, usually same day. No sales call unless you want one.
We will come back with a plan and an hour estimate, usually same day. No sales call unless you want one.
From $30/hour · 1-language minimum · human native testers · reply usually same day