Phrase (Phrase TMS, Phrase Strings, Orchestrator, Language AI)

Orchestrator explicitly combines triggers and actions into branching workflows and routes QPS-scored content to post-editing or LQA.

Public TMS APIs create projects, attach TM/TB/engines, upload jobs, analyze and pretranslate, with asynchronous callbacks for integration.

Platform-level evidence explicitly routes multiple engines and language models by language pair and content type; this establishes the C001 suite context directly.

TMS connectors pull and plugins push content between CMS/marketing systems and translation; connector availability depends on subscription or add-on.

Strings imports and exports localization resources and integrates with CI/CD, establishing software delivery; detailed encoding and placeholder preservation at each integration boundary remain unverified.

Strings supplies contextual editing, Figma previews and UI constraints that give translators direct visual and length context.

The memory overview establishes shared project memories and overwrite/update behavior; dedicated stale-entry cleaning and administrative search were not verified.

Concept-based multilingual termbases hold approved terms, assign to multiple projects and feed translation-editor suggestions and terminology QA.

The platform explicitly brings memory, terminology and style-guide context into model-based translation; this supports grounded first drafts.

TMS stores confirmed segments and retrieves matching prior translations; shared memories support consistent reuse across projects.

Strings contextual editing and Figma previews explicitly expose localized copy in visual context before delivery.

QA compares source and target, checks terminology, tags and spelling, and can require warnings to be resolved before completion.

The OTA documentation explicitly delivers translation updates independently of app releases, subject to SDK integration and usage quotas.

Strings explicitly syncs keys through API and GitHub/GitLab CI pipelines, keeping product strings and translations aligned during development.

The TMS API describes the CAT editor and assigned TM/TB, while QA and term suggestions support professional editing; predictive typing and all filter guarantees are not established.