Skip to content
Last updated

September 2025 Release Note

These are the Treasure Data features and improvements released in September, 2025.

Audience Agent Update - Segment Analytics & Existing Segment Utilization

  • This update enhances the Audience Agent by enabling marketers to analyze, refine, and build new segments using existing segment assets. For example, a marketer may want to add new conditions to an existing segment to easily refine target audiences.

  • Marketing analysts can now review segment metrics, obtain human-readable logic explanations, and create derivative segments for more efficient and precise audience targeting.

  • Learn more by visiting the Treasure Data Product Documentation site.


Agent Context: Knowledge Base for Parent Segment & Attributes

  • Agents can now read Parent Segment configuration metadata, folders, and segment data for analytics and recommendations.

  • Marketers and prompt engineers benefit from smarter, context-driven insights, intuitive data labels, and improved analysis workflows when working with segmented audiences.

  • The Knowledge Base is created automatically once the LLM is enabled on a particular parent segment, and marketers get smarter answers from their agents since the agent “knows” segment structure and definitions, so it can explain the ‘why’ behind segment logic.

  • Learn more by visiting the Treasure Data Product Documentation site.


Image Upload Option in Foundry Workspace

  • This feature allows users to upload images (including .png and .jpeg) directly into forms within the Foundry Workspace, enabling image submissions as part of content generation workflows. Agents can also lightly modify a provided image, such as changing the background.

  • Designed for marketers and designers who need to include images in their AI-driven content creation or conversational agent interactions, this enhancement streamlines workflows that require visual inputs.

  • Learn more by visiting the Treasure Data Product Documentation site.


Output History in Foundry Workspace

  • Users can now view, manage, and resume work on previously generated output within the Foundry Workspace, supporting both form and chat-based workflows.

  • Users, including marketers and content creators, can access recent outputs, revisit drafts, and pick up tasks where they left off, improving productivity and supporting iterative content generation (i.e. create a new version of a Welcome Email for new subscribers).

  • Learn more by visiting the Treasure Data Product Documentation site.


Improved Workflow Execution: Agent-Triggered Automation

  • This feature allows agents to automatically execute selected workflows with specified parameters, simplifying tasks that require data preparation.

  • Prompt engineers and automation developers can now integrate workflow execution into agent-driven processes, providing secure and efficient automation for their use cases.

  • Learn more by visiting the Treasure Data Product Documentation site.


Claude Sonnet 4: Now on Tokyo Region

  • Claude Sonnet 4 (normal mode) to Production is now available in the US and Tokyo regions.

  • Claude Sonnet 4 means agents are using newer Claude models - specifically, Claude Sonnet 4 is considered better than previous models (like Sonnet 3.5) due to its improved reasoning, superior coding quality, enhanced contextual understanding, and faster performance, particularly for complex coding tasks and large-scale codebase navigation.

  • Learn more by visiting the Treasure Data Product Documentation site.


New: Teradata Vantage JDBC Driver (input)

  • This enables customers to import data from Teradata Cloud via table, view, or query with incremental support.

  • The connector leverages the official Teradata JDBC driver and supports username/password authentication, incremental loads (Long, Timestamp), and error handling aligned with JDBC.

  • Output connector (write-back to Teradata) is planned for a future phase.

  • Learn more by visiting the Treasure Data Product Documentation site.


Taboola API Connector - First Party Audience API

  • We’ve released the new Taboola Bulk Export Connector supporting the First Party Audience API.

  • Marketers can now create First Party Audiences and Add/Remove users via EMAIL_ID or DEVICE_ID directly from Treasure Data.

  • This release includes OAuth authentication, audience auto-creation, and scalable batch processing.

  • Learn more by visiting the Treasure Data Product Documentation site.


Dynamic Dropdown - Iterable List

  • We’ve enhanced the Iterable Export Connector to support an auto-populated, searchable dropdown on Activation UI for List IDs using Iterable’s Get Lists API.

  • Marketers no longer need to manually copy/paste List IDs when using Add/Remove operations, reducing errors and setup time.

  • This change improves reliability for Iterable Activations.

  • Learn more by visiting the Treasure Data Product Documentation site.


Data Connector Meta Custom Audience (Kaizen): Support a logic to handle session_id

  • We’ve added support logic for Meta Replace API session parameters in the Meta Custom Audience connector.

  • This ensures safer multi-batch audience replacements within the 90-minute session window.

  • Customers can now reliably manage large audience uploads without manual workarounds or timeout risks.

  • Learn more by visiting the Treasure Data Product Documentation site.


Kaizen: [HubSpot Input] Update the OAuth scope to support custom object

  • We updated the Hubspot Authentication OAuth scope to support custom objects. With this change, customers are able to ingest records from Hubspot Custom Object.

  • If you already have Authentication with Hubspot, please re-authenticate the setting to grant the new scope.

  • Learn more by visiting the Treasure Data Product Documentation site.


Kaizen: Adobe Target Bulk Profile Export - v1 support

  • We’ve added support for selecting the Adobe Target Bulk Update API version (v1 or v2) in export settings.

  • This allows marketers to align API behavior with their implementation - v1 for pc_id update-only, or v2 for create-if-missing/third_party_id.

  • Existing connector jobs will default to v2, while new exports can explicitly choose the version.

  • Learn more by visiting the Treasure Data Product Documentation site.


Expanded S3 Workflow Operator Support with AWS AssumeRole

  • This release adds new S3 workflow operators (s3_copy, s3_move, s3_delete) and support for ‘s3_get’ and ‘s3_put’ in custom Python scripts, all utilizing AWS AssumeRole for secure authentication.

  • Data Engineers can now automate S3 operations within Treasure Data workflows while meeting security and compliance requirements by avoiding direct management of AWS access keys.

  • For more information, refer to the Treasure Data Product Documentation site.


Real-time A/B Testing Engine for Trigger Journeys

  • This feature enables marketers to create, modify, and manage real-time A/B tests, assigning users to test variants and tracking variant membership for campaign optimization.

  • Designed for marketers running Trigger Journey workflows, it allows incremental variant assignment to new profiles.

  • Learn more by visiting the Treasure Data Product Documentation site.


Image Editing Agent: Background Removal and Variation Creation

  • The Image Editing Agent is now available in the Creative AI Suite.

  • This AI Agent includes both background removal and image variation capabilities in a single, streamlined workflow.

  • Learn more by visiting the Treasure Data Product Documentation site.