Controlled by your team
- Brand Voice
- Terminology
- Approved Claims
- Forbidden Claims
- Market Rules
- Previous Approved Language
- Content Type Defaults
These only change when your team decides they should.
AI output does not rewrite these. A reviewer edit does not silently become a rule. Localizethat may surface a suggestion based on approved work, but a suggestion has no effect until someone accepts it.
Grows from approved work
- Translation Memory
Finalized approved translations can become reusable language for future localization. This is approved language memory — not model training.
Some context grows. Your rules stay controlled.
A large context library is not a large prompt
A large Context library does not mean Localizethat sends everything into every localization. Localizethat selects the context relevant to the content, language and market you are working with.
That means you can keep growing Company Context without every localization becoming noisier.
Where to start
You do not need to complete Company Context before your first localization. Start with the context that matters most to the content you localize the most, and build from there.