Software localization is the work of preparing an app or website so it feels native in every language you support, from button labels to date formats. Done by hand, it means exporting files, emailing them to translators and merging the results back, over and over. This tutorial shows a cleaner path, using Wordbeam to connect your repository, keep strings in sync and ship translations at the speed of your commits.
It is written for developers and product people who are localizing for the first time, so each step explains the idea before the tool.
Step 1: understand what needs to be localized
Before touching any tool, list what users actually see. Interface text is only the start. Dates, numbers, currencies, plural forms, images with text and error messages all change by market. The standard terms are explained in the overview of internationalization and localization: internationalization prepares the code, and localization adapts it for each locale.
Internationalize once, localize many times. If strings are hard coded in the interface, move them into resource files first, because every localization tool depends on that step.
Step 2: pick a localization workflow that matches your pipeline
A good localization workflow follows your code instead of fighting it. New keys should appear for translators when developers commit, and finished translations should return as normal pull requests. Anything that needs manual file handling will eventually be skipped, and then your releases ship with missing text.
Frameworks matter too. If you build with React and server rendering, this guide to Next.js i18n shows how routing and message files fit together before you connect a platform.
Step 3: connect your repository to Wordbeam
The software localization platform in Wordbeam is built around repository sync. You connect a repository once, and the system keeps strings, branches and releases aligned. Engineers never export a file by hand, and translators never touch a line of code. The features that matter most are:
- Two way sync. New keys flow in on every commit, and confirmed translations return as a clean pull request.
- CLI and REST API. Trigger pushes, pulls and status checks from CI with one command.
- Branch processing. Each feature branch gets its own scope and merges back when it ships.
- Placeholder and ICU protection. Variables, ICU plurals and HTML tags are locked as non editable tokens, so a translator cannot break a build.
- Many formats. More than 16 formats, including JSON, XLIFF, PO, DOCX, XLSX, IDML and SRT.
Native connectors for GitHub and GitLab handle the repository side, and the same platform also connects with Figma, Contentful and AEM when your content lives elsewhere.
Step 4: translate with memory, glossary and AI
Once strings arrive, translators work in an editor with translation memory and a glossary, so repeated labels such as Save, Cancel and Settings are translated once and reused. For first drafts you can pre-translate with GPT, Claude, Grok or DeepL, then have a person review. Wordbeam lists more than 100 languages, which covers most product roadmaps.
| Stage | Who does it | Tool support |
|---|---|---|
| Extract strings | Developer | Repository sync |
| First draft | AI engine | Multi engine pre-translation |
| Review and edit | Translator | Memory, glossary, QA checks |
| Return to code | Automatic | Pull request |
| Release | Developer | CI check |
Step 5: automate the loop in CI
The last step is to remove yourself from the process. Add a CI job that pushes new keys on each merge and pulls approved translations before a release. This is often called continuous localization, and it keeps the gap between a new feature and its translated version close to zero. Wordbeam lists the time from commit to editor at under 30 seconds, so translators see new text almost at once.
Test with a pseudo language or a long language early. If a button still fits when the text grows by thirty percent, it will probably fit in most real translations, and you will find layout problems before users do.
Common mistakes to avoid
- Concatenating strings in code, which breaks grammar in other languages.
- Leaving out context, so translators guess what a short label means.
- Ignoring text expansion, since German or French labels are often longer than English ones.
- Skipping a final check in the running app.
Pricing and free trial
You can test the whole flow in a 14 day free trial that includes 10,000 words. After that, the Freelancer plan costs $25 per month for 100k words, and the Studio plan costs $95 per month for 300k words.
Frequently asked questions
How is app localization different from translation?
Translation converts words. Localization also covers formats, layout, plurals and cultural fit, and it needs a technical workflow around the code.
Will translators break my code?
Not if variables and tags are protected. Wordbeam locks placeholders, ICU plurals and HTML tags as tokens that cannot be edited.
When should I start localizing?
As early as possible, because moving strings into resource files is far easier before the product grows.
Localization stops being a project when it becomes part of your pipeline. Connect one repository, push a branch and watch the strings appear in the editor.