PTC
Sign Up Upgrade to Pro Your Account

PTC Private Translation Cloud: AI localization for software

PTC (Private Translation Cloud) understands your app, what each screen is for and what each text does, and translates it to read native in every language.

Start translating your app for free Your Account Upgrade to Pro Your Account
  • Set up through MCP by your coding agent
  • Publish-ready AI translations, reviewed before delivery
  • Pay as you go, per word. No seats, no fixed costs.

Connect your coding agent, get a localized app

Connect PTC to Claude Code, Codex or Cursor through MCP. PTC takes your app from where it is to fully localized, and delivers the translations as a pull request you approve. You decide the languages, the price and the delivery once.

1

Your agent prepares the app. It finds every text, fixes code that translates badly and sets up the files.

2

PTC translates and reviews every language. It uses your terminology and works while you do other things.

3

PTC delivers publish-ready translations into your project. You ship in languages you do not read.

How the PTC agent guide works · PTC MCP server reference

Your agent prepares your repository, PTC returns the translated files reviewed, and a pull request delivers themLeft, your repository with its English locale file and one code line fixed by the agent. Right, PTC's list of translated files, one per language with its flag and a review check, continuing downward. Below, a pull request card in your project with an Approve button.Your repositoryClaude CodeCodexsrc/app.tsxlocales/en.json- t("Hello " + name)+ t("hello", { name })// one text, not three piecesde.jsonfr.jsonja.jsones.json………translated and reviewedPull request from PTC4 languages added, 4 files changedApprove
Your glossary, your code and your website build one glossary and brand note, and the same term appears on every screenLeft, three sources, your glossary, your code and your website, feed one glossary card with three terms. A brand voice card joins it. Right, three app screens in three languages show the same term on the same button.Your glossaryYour codeYour websiteGlossaryWorkspacekeepPlanTarifSyncSynchronisierenBrand voicedirect, for developersWorkspace erstellenCréer un WorkspaceWorkspace を作成

Consistent terminology and your brand voice, in every language

PTC settles your terminology once and uses it in every text, so a feature has one name on every screen. It also follows your brand voice, so your product sounds like your product in every language.

  • A glossary that keeps your terms and your brand consistent

    If you have a glossary, PTC starts with it. If not, PTC builds one from your code, your website and its own research, and asks you only where your sources disagree. How the glossary works.

  • Your tone and audience, in every language

    Describe your product and its readers once. Every translation follows. Translation for your brand and audience.

  • Texts that fit the screen they sit on

    Your agent reads the code around an unclear text. PTC turns it into a note that guides the translation in every language.

Publish-ready translation quality

You can ship a language you do not read. Every translation is measured, compared and checked before delivery, and covered by the 60-day Translation Quality Guarantee.

  • Scores above DeepL and human translators on our benchmark

  • Better than a script that calls ChatGPT or Claude

    A script has no glossary, no context and no check, so one term comes back three ways. PTC compared with DIY scripts.

  • Every language reviewed before delivery

    PTC reviews each translation against your description, glossary and notes, and fixes what it finds.

  • Your translated screens checked visually

    With AI Visual QA enabled, PTC reads screenshots of your translated screens and flags what looks wrong. AI Visual QA.

Quality shown four ways: the benchmark ranking, one term kept consistent, every language reviewed, and a translated screen fixedTop left, our benchmark's ranking with PTC first, human translators second and DeepL third. Top right, a script that calls an LLM returns one term three ways while PTC returns it once. Bottom left, four language flags each with a review check. Bottom right, a translated screen with a button whose text overflows, marked in red, and the same screen with the text fixed, marked in green.Our benchmark1PTC2Human translators3DeepLOne term, three languagesA script that calls an LLMArbeitsbereichWorkspaceArbeitsflächePTCWorkspaceWorkspaceWorkspaceReviewed before deliveryAI Visual QAEinstellungen speichSpeichern
The loop: a push, PTC translates the change, reviews every language, and a pull request comes back; the next change starts it again and only the changed words are billedFour steps in a loop: you push a change, PTC translates only the new texts, PTC reviews every language, your CI opens a pull request you merge. A dashed arrow returns to the first step labelled next change, and a badge reads billed: 12 words.You push a changeto your default branchPTC translatesonly the new and changed textsPTC reviewsevery languageYour CI opens a pull requestyou merge itnext changebilled: 12 words

Set up PTC once, and your app stays fully translated

After the setup, every change you push comes back translated and reviewed as a pull request. You pay only for the words that changed.

  • Every release ships in every language

    Your agent commits the one file GitHub or GitLab needs. From then on, you only merge pull requests. Git integration.

  • You pay for the changes, never twice for the same text

    Unchanged texts stay as they are. A one-line fix costs a few words.

Try it free on your own app

The trial covers 20,000 words into 2 languages, with no credit card. After that, pay as you go, per word. No subscription, no seats, no minimum.

Pricing

Start your free trial