Judge copy in the layout it ships
Visual, strings, and code modes share one source of truth so marketers and linguists stay aligned.
3.1 Translate
Translate the page itself - layout, structure, and tracking stay intact in every market.
Visual, strings, and code modes share one source of truth so marketers and linguists stay aligned.
Pending, approved, and published stay explicit - with staleness flags when the source quietly moves on.
1.0
Source on the left, target on the right, in the layout it will actually ship in. Image text is re-rendered in place, not just captioned.
3.2 AI image translationTranslation runs against the page itself, not an exported spreadsheet, so layout and tracking stay intact.
When source copy moves, only the affected strings are flagged. Last approved versions keep serving.
2.0
The same content, presented the way each role needs to see it - with coverage visible at a glance.
Edit copy on the rendered page.
Segments, limits, glossary.
Markup with nodes marked.
Gleiches Layout. Gleiches Tracking. Text in der Sprache, die konvertiert.
Gratis testen3.0
A three-state workflow with an explicit machine-translation marker, plus staleness detection when the source quietly moves on.
Reviewers sign off before live. Approved copy is locked against accidental regeneration.
Changed strings are marked stale per market. Live translated pages keep serving the last approved version.
AI output waiting on a market reviewer. Tagged AI translated throughout.
MT draftA reviewer signed off. Locked against regeneration.
LockedServing live on the language URL, with hreflang in place.
Live4.0
Anything a visitor or a crawler can read - across twenty-four languages supported today.
Headlines, paragraphs, lists, and captions.
Same structure. Same tracking.
Start free trialGleiches Layout. Gleiches Tracking.
Gratis testenShips in the rendered layout, not a spreadsheet cell.
FAQ
For structural and functional copy, yes. For high-stakes headline and offer language, treat AI output as a first draft: it lands in review, not in production, unless you explicitly enable auto-publish.
Yes, in any of the three review modes. Edits are versioned and are never overwritten by a later regeneration unless you ask for it.
Most teams keep a reviewer per market rather than a translator. The work shifts from producing text to approving it, which is roughly an order of magnitude less time.
Yes. Arabic and Hebrew flip layout direction, mirror padding and alignment, and keep numerals and Latin brand names in the correct direction.
Yes. Each language gets its own slug, which you can override per market rather than accepting a transliteration.
Automatically. Every language version links to every other, with a self-referencing tag and an x-default, regenerated whenever you add a market.
Affected strings in every language are flagged stale with an amber banner. The live translated page keeps serving the last approved version until you review, so nothing breaks mid-campaign.
Yes. Import a TMX file per language and matched segments are reused verbatim before the model is called, which also reduces usage cost.
Free for 14 days, no card required. Larger teams can book a walkthrough instead.