VidiFlow Release Notes

The following items on the list encompass breaking changes, features, and fixes that are relevant for the major release.

Visit this page to get an overview for all releases and patches.

Patches for this major release

Patches are shown based on recency and will appear below as soon as they are issued.

Release 25.4.1 on 26 Jan 2026

Breaking Changes

Setup and Operations

The following central components were updated and will be shipped in this version, along with relevant product and service dependencies:

Please note that the UPDATED marker indicates a change between the previous major release and the one this document is issued for.

Features

New Tasks

The task AssociateFileToShapeComponents (1.0) associates a file given by a VidiCore file id to existing shape components. The file has to be an Op1a file. The caller has to make sure that the file object's asset is binarily identical to the asset already present in that shape. This cannot be verified by the agent and not even by VC since the file may be located on an inaccessible storage.

NormalizeMxf (1.0) uses the VidiCore VidiCoder to normalize an MXF file. This can be used instead of NormalizeMxfViaLinux to use VidiCoder instead of the MediaFramework.

To create index file for faster navigation for proxy playback, the new task CreateProxyIndex (1.0) can be used. It uses VidiCoder instead of MediaFramework.

The task pair GetMetadataJson (1.0) and UpdateMetadataJson (1.0) can be used to query or set metadata in JSON format from VidiCore. The tasks can be used equivalently to the GetMetadata and UpdateMetadata tasks, which use XML format.

Two new cognitive service tasks DeepVAMultiModelAnalyzer (1.0) and AwsTranscribeAnalyzer (1.0) allow to run different AI services on VidiCore shapes to enrich item metadata automatically:

• DeepVAMultiModelAnalyzer enables simultaneous execution of multiple DeepVA mining modules, such as face, object, and landmark recognition, as well as speech-to-text (STT) and specialized tasks like detecting lower thirds. For a full list of available mining modules, visit DeepVA’s documentation. The integration supports cloud-only, hybrid, and on-premise deployments - please reach out to us for further details.

• AwsTranscribeAnalyzer leverages the AWS Transcribe service to enrich VidiCore metadata by adding speech-to-text (STT) transcripts. To run the analysis, MP4 files must be stored in an S3 bucket, and the appropriate AWS IAM permissions must be configured to allow the service to access and process the content.

Updated Tasks

A new boolean parameter named ChunkedWrite is introduced in some tasks. It allows one to enable the chunked writing of the target file to blob storages like S3. The following tasks are affected:

• TranscodeProxy (3.3)

• CopyFileVidispine (2.5)

• RenderItemSequence (4.2)

• TrimVsa (1.1)

Resource Tags

The usage of resource tags was enhanced to use inherited resource tags for storages.

Workflow Monitor

Allow Priority Change on Rerun of Workflows

If a user wants to rerun failed workflows in Workflow Monitor, it is now possible to change the priority for the rerun. This allows one to control if these workflows are executed sooner (high priority) or later (low priority).

Other improvement

Introduce Housekeeping config as environment dependent

The following values in the housekeeping configuration are now environment dependent:

• DeletionLock duration

• Recycle Bin retention time

This allows one to configure different values in production and test environments.

Fixes