Content Groups
Structure your content into groups, each with its own context and guidelines.
It starts with your style guide. Content Groups turn your brand’s rules into a contextual layer built for each type of content, shaping every translation from generation to quality evaluation to AI post-editing. The result is tailored translations at scale.
A pipeline for every content type
Content Groups turn your guidelines into structured context that drives content-specific, automated translation pipelines.
01 Rules
A style guide your AI can act on
Start with your style guide and let Phrase split it into individual rules, each one structured so AI can apply it reliably. Instead of asking the AI to interpret a whole document at once, every rule becomes clear, focused guidance it can follow consistently across your content. And when you need something your style guide doesn’t cover, you can add your own rules and checks anytime, shaping the guidance exactly the way you want.
02 Content Groups
The context each content type needs
Bundle the rules that matter into Content Groups, one for each type of content, from marketing to legal to product. Each group becomes the contextual foundation for its own automated translation pipeline, giving every content type the specific guidance it needs. Because rules aren’t locked to a single group, you can reuse them across groups and languages, so every rule you create once can be applied anywhere it’s relevant.
03 AI Checks
Evaluate against the exact same context
Turn the same rules into AI Checks, so quality evaluation runs on the exact context that guided translation in the first place. There’s no gap between how content is created and how it’s assessed, because both draw from the same source. Every content type gets automated, criteria-based QA that catches issues before they ship, and because the rules are already in place, there’s no separate setup to manage.
What Content Groups unlock
The same context flows through every stage, so automation scales without giving up control.
Bespoke translations at scale
Every content type runs on its own tuned pipeline, so output matches intent automatically.
Dramatically less post-editing
You get what you are aiming for the first time, cutting manual review and rework.
Context across every stage
The same context drives translation generation, quality evaluation, and AI post-editing.
Reusable and governed
Define rules once and reuse them across content groups and languages, from a single source of truth.