Verison 3.6

Version 3.5

17 July, 2025

Enhancements

VA

• Global Entity Versioning:
  Global entity versioning is now supported for VA projects. After the first publish, each project gets its own versioned global entity, and no migration is required. Draft and published versions are separated — changes made in the UI affect only the draft version. The published version keeps the global entity values as they were at the time of publishing.

Improvements

VA

• New Settings for Generative Q&A
  New configuration options have been added to the Generative Q&A Settings page, allowing users to control extraction strategy, embedding dimensions, crawl depth, and max pages to crawl. The “Retrieved Document Split Count” limit is now increased to 20. New fields like Maximum Pages to Crawl, Content Extraction Strategy, and Maximum Crawl Depth are configurable and mapped to the training settings API accordingly. These settings are also passed as environment variables to the GenQ&A pod and are supported starting from metadata version v21.

  • Adding "Tenant ID" Info to Sestek TTS Settings
    A new optional field called ABBR List ID has been added to the Sestek TTS integration settings. This allows VA users and BACs to include a custom abbreviation list by passing a Tenant ID in the TTS API request. When filled, the value is sent as the "Tenant" header in the backend TTS request. A tooltip has been added next to the field for guidance.

• Minor UI Improvements:
  Several minor improvements have been made to the user interface to enhance user experience and usability.

• Minor Bug Fixes:
  Various minor bugs have been addressed to improve overall system stability and performance.