New APIs, style guides, and agent-enabled infrastructure make linguistic intelligence accessible across enterprise workflows.
Phrase, the world’s leading linguistic intelligence platform, today announced a series of new platform enhancements designed to help organizations extend linguistic intelligence across human and agent-based workflows.
Winning in the global market is never about churning out content. It’s about creating trusted, consistent, and on-brand content in every language. ”
— Georg Ell, Phrase CEO
As AI agents become more active participants in content workflows, organizations face new challenges. Systems that generate content often operate without access to the language assets, brand standards, and governance frameworks needed to create consistent customer experiences across international markets.
This enhancement reflects a broader shift to global content operations where humans and AI agents increasingly work together. Phrase directly addresses this problem by making context, quality control, and language assets accessible across workflows, enterprise applications, and the agents running within them. The result is greater consistency, governance, quality, and scaling as content creation increases.
Also read: AiThority interview with Matej Bukovinski, Chief Technology Officer at Nutrient
“Our customers have spent years building linguistic context, including glossaries, translation memories, quality standards, and brand voice. With the now publicly available style guides and quality profiles, and the new API we’ve published (MCP), that context can power every agent, workflow, and system across the enterprise. It’s the difference between AI that just generates output and AI that generates content that actually works in the marketplace.”
A platform built for both humans and agents
Organizations increasingly want the freedom to use their own models, build their own agents, and integrate linguistic intelligence into the systems and agent tools they already use. To support this, Phrase has expanded its API portfolio and significantly expanded its Model Context Protocol (MCP) capabilities.
New APIs for style guides, quality profiles, and language assets give agents, workflows, and enterprise applications access to the same brand context, terminology, translation memory, and quality ratings that power Phrase’s unique AI capabilities.
Expanded MCP support includes features such as translation memory and termbase management, quality assessment, analysis, and citation capabilities. You’ll also be able to manage your files from MCP-compatible AI tools and agents, such as Claude Desktop and the Google Drive connector. These enhancements make Phrase more open, flexible, and accessible across enterprise applications, custom AI solutions, and agent-based workflows.
Governance built for automated content operations
This update also makes the style guide generally available.
First introduced earlier this year, style guides provide both AI and linguists with the context behind the content, brand voice, tone, audience guidelines, and market-specific requirements that determine not just what to translate, but how to translate it. They are applied consistently throughout the localization process, from initial translation generation to post-editing and quality assessment, so all stages work from the same contextual foundation. Open API access means any agent, workflow, or external system across your enterprise can access the same foundation. This ensures that the context that defines how brands communicate in any market is no longer limited to the platform, but available wherever content is created or adapted.
Quality profiles available through API access allow organizations to evaluate content based on the same criteria they used to create it.
Together, they go beyond traditional “quality” measures (such as MQM) to establish a shared source of truth for translation, evaluation, and adaptation, helping organizations improve quality, maintain consistency, and scale content operations globally with confidence.
Zero-touch localization for instant translation of every release
Phrase is introducing zero-touch localization, a new automation that eliminates the manual effort that typically exists between code changes and translation releases. Once configured, the pipeline automatically runs in parallel with your existing development workflow. Detect new content, trigger translations, and return completed strings to your codebase without requiring any action from developers or localization teams. This reduces engineering overhead and accelerates release cycles. GitHub business and enterprise customers can request early access now.
Additional platform enhancements
Phrase Studio: The latest release extends Phrase Studio to bring the same translation memories, style guides, and quality standards used across global content operations to video localization workflows. New features include support for files up to 20 GB, automatic subtitle timing suggestions, undo and redo functionality, and enhanced dubbing workflows.
Design-driven localization: Enhancements to the Phrase Figma plugin improve performance and automation for design teams managing multilingual experiences with support for large design systems, improved content synchronization, and branch-aware automation.
Expanding ecosystem connectivity: Through our partnership with Blackbird, Phrase now connects to more than 200 enterprise applications and adds 150 new integrations including Notion, ServiceNow, Airtable, Asana, and Bynder, helping organizations unify linguistic intelligence across their existing technology stack.
Foundation for the future of global content
As organizations increasingly rely on AI to create, adapt, and distribute content across markets, the need for governance, context, and consistency continues to grow.
Frese believes the future lies in a shared intelligence layer that connects people, systems, and agents through a common understanding of language. These updates expand on that vision and help organizations bring context, quality, and governance to workflows where humans and AI increasingly work together.
“The most common question we hear from customers today is whether they can build their own agents, their own models, and their own workflows without sacrificing the quality and consistency that defines their brand across markets,” Elle said. “Winning in the global marketplace is never about producing tons of content. It’s about producing trusted, consistent, and on-brand content in every language. Our language intelligence platform is designed to help you do that at scale, and to get measurably better the more you use it.”
Also read: AI Systems – Interoperable AI Systems: Connecting models across platforms
[To share your insights with us, please write to psen@itechseries.com ]
