Run every client project
from one workspace

Yes — Team and Business plans include unlimited projects and unlimited seats under one organization, so each client gets an isolated project without a separate subscription.

Batch AI translation drafts a whole set of untranslated keys at once with Claude, preserving placeholders and tokens, so your linguists start from a first pass instead of an empty cell — and translation memory means repeated phrases across clients don’t cost quota twice.

Invite clients or stakeholders with an appropriate role and they can see live translation states — AI-translated, human-edited, reviewed — directly in the project.

Pricing is per organization, not per client project — Team and Business plans include unlimited projects, so growing your client roster doesn’t mean a new bill for every engagement.

Yes — every project gets its own glossary, and terms only apply within that project unless you deliberately link a shared org-wide glossary to it. That keeps one client’s brand terms from leaking into another’s translations, while still letting you share common terminology where it makes sense.

Projects, glossaries, keys, and translations are isolated per project — a glossary term never leaks between clients. Translation memory is the deliberate exception: it is a single global, anonymous pool keyed by source hash, locale, and model, with no project or organization attached. That is why boilerplate reuse works across engagements at no quota cost, but it also means an exact-match string can cross a client boundary. If a client’s contract rules that out, turn translation memory off for their project and it will neither read from nor write to the pool.

Translation QA (Team and up) scores every translation and surfaces categorized findings, and you can export a per-language MQM scorecard as a concrete quality artifact — score trend, top finding categories, and share of rows needing review — so a client sees a real signal, not just "we reviewed it," before final delivery.

Yes. The external-vendor workflow lets you assign specific locales or keys to an outside translator, track each assignment from open through submitted, and gate approval — with a two-person split so the assignee and reviewer are never the same person. It is built on the same per-language grants and notifications, so an external linguist only ever sees the scope you assigned them.

Yes — a per-project AI Profile (Team and up) captures each client’s brand voice, tone, and worked examples, and grounds every AI translation for that project in them, alongside that project’s glossary. Because it is per project, one client’s voice never bleeds into another’s.