Latest Releases
Manual Error Handling Improvements
We’ve updated the default manual error handling behavior settings for new workflows.
What’s changed?
Legacy Manual error handling is now disabled by default for newly created workflows. Please see your workflow settings to enable this.
Existing workflows will continue using their current manual error behaviors unless you choose to update them in the workflow settings.
For details on how to configure error handling, visit our What is an Error documentation.
Vector Tables updated
The Vector Tables connector has been updated with the following:
New operations:
Bulk Update Vectors by Query
TikTok Business updated
The TikTok Business connector has been updated with the following:
Operation updates:
Added error handling for token request
Updated Zendesk OAuth authentication flow
We’ve updated the OAuth authentication journey for the Zendesk and Zendesk Trigger connector versions to resolve issues with creating new authentications - no action is required on your part.
What has changed?
The OAuth popup flow has changed slightly: after completing third-party consent, users will see a new intermediate screen before returning to the app.
We have disabled editing existing Zendesk OAuth authentications from within the embedded config wizard .
How are you affected?
No action needed - existing authentications continue to work as normal.
Only the OAuth UI authentication journey is affected.
Embedded: For end users to make changes to an existing OAuth authentication (e.g. update scopes) via the config wizard, they'll need to create a new authentication instead of editing the existing one.
If you use custom CSS for embedded, you may want to review it in case adjustments are needed for the updated screen.
File Storage System: Tray Native File Storage
We’ve introduced the File System, a fully managed file storage layer within Tray. It allows you to upload and store files directly in the Tray platform, then access them in your automations using the File System connector.
This unlocks critical capabilities for building AI Agents and advanced automations that rely on handling unstructured data - such as PDFs, HTML pages, transcripts, or other documents - retrieved from external systems. Instead of sending data to third-party storage, you can now keep everything within Tray, simplifying architecture and reducing integration overhead.
What this enables:
Store and retrieve files as part of your projects
Build AI Agents that use file-based data sources (e.g. PDFs or transcripts for context)
Parse, enrich, or summarize file contents using AI
Learn more in our File Storage System documentation.
Note: The File System feature is not supported in Embedded solutions.
Merlin IDP (Beta) updated
The Merlin IDP (Beta) connector has been updated with the following:
New operations:
Extract Text (image/pdf)
Extract text structure as HTML (image/pdf)
Extract text structure as Markdown (image/pdf)
AWS Bedrock updated
The AWS Bedrock connector has been updated with the following:
New operations:
AI Generation
Create emebeddings
List models
Raw HTTP
Authentication updates:
Added support for session_token authentication.
File System updated
The File System connector has been updated with the following:
Operation bugs:
Fixed default value for operations that uses File ID