Smarter MT updates, built directly into your Trados workflow

If you use Trados and Language Weaver, you already know how powerful AI translation can be. The challenge is that updating models hasn’t always kept pace with how fast localization teams actually work. 

That’s exactly what the new Auto-Adaptive Model Updater App is designed to solve. Now available in the RWS AppStore, this app creates a seamless, fully automated link between Trados workflows and Language Weaver auto-adaptive model training. This integration enables an industry-unique level of automation, allowing your MT to keep improving, project after project, with no extra steps for your team after the initial setup.

For many Trados and Language Weaver users, updating auto-adaptive models has traditionally meant stepping out of their workflow, exporting TMX files and manually importing them into Language Weaver. It works, but it takes time and breaks momentum – especially when projects are moving fast and deadlines are tight. And because it is a separate step, it can sometimes happen less often than intended.

The new app removes that friction by simplifying model adaptation.  

It introduces a new task inside Trados custom workflows, designed for customers using Auto-Adaptive or Generative Language Pairs. This allows Language Weaver models to learn continuously from your approved translation project content, without manual intervention. 

Once added after the “Update Translation Memory” step, the process runs seamlessly in the background. Edited segments are automatically collected based on your chosen fuzzy match threshold and sent securely to the relevant Language Weaver auto-adaptive model. When Language Weaver retrains that model, the latest project content is now already included. 

There’s no need to download files, upload them elsewhere or manage retraining schedules. The work your team has already approved is simply reused at the right moment, without disrupting established processes.   

Security and data residency are maintained end to end. Files stay within the RWS suite, with secure, encrypted data transfer between Trados and Language Weaver, ensuring compliance with your data security requirements.

By creating a self-reinforcing learning loop between Trados and Language Weaver, each completed project helps improve the next one. 

For localization teams, that means less admin and better results at the same time. Models adapt faster to your translation, domains and terminology. Project managers stay inside the tools they already use. And your investment in Trados and Language Weaver delivers more value over time. 

Just as importantly, it removes a common point of delay. Model updates no longer depend on someone remembering to take manual action after a project closes. 

For corporate translation teams and LSPs managing high volumes of content, this matters. These teams don’t need another system to manage – they need their existing systems to work better together. By bringing model adaptation directly into Trados, we make MT improvement part of the production workflow, not a separate task to fit in later. And because this is a step in a custom workflow, it means that you choose which data updates the model per workflow, content type, project type, line of business, etc., while still providing you with full control over your data.

Talk to an RWS expert about using Language Weaver and Trados together to keep your localization moving faster – and learning continuously.