Note
Access to this page requires authorization. You can try signing in or changing directories.
Access to this page requires authorization. You can try changing directories.
This article lists release notes for Azure AI Video Indexer enabled by Azure Arc. For cloud-based release notes, see Azure AI Video Indexer release notes.
August 2026
Multi-GPU scaling for real-time analysis
You can now configure multiple GPUs for real-time analysis in a single Azure AI Video Indexer extension. When you add more GPUs, you add more real-time analysis instances and increase the aggregate camera capacity. Deployments that use one GPU continue to work as before.
Select GPU count when you create an extension in the Azure portal, or configure the GPU capacity by using Helm values. If no GPU has enough remaining capacity, a camera enters the Queued state and starts automatically when capacity becomes available.
For extension setup details, see Manage real-time analysis extensions. For planning guidance and per-GPU camera benchmarks, see GPU capacity and queued cameras.
June 2026
Video Agents Foundry solution
Video Agents Foundry solution is an end-to-end platform that helps businesses automate complex video analysis by using AI agents. It combines Azure AI Video Indexer, Azure Arc, and intelligent agents to deliver real-time insights, automated decision-making, and scalable edge-based video analytics.
The solution is available in the Microsoft Foundry solution templates catalog and as an open-source sample on GitHub.
March 2026
Custom AI insights for objects with attributes
Create object insights that match your specific operational needs. Describe objects with attributes in natural language: specify colors, types, conditions, or combinations like "a red classic car" or "person wearing a hard hat." Add example images and negative examples to refine detections and improve accuracy.
For more information, see Custom insights.
January 2026
Situation custom insight
Situation custom insight enables you to define scenario-based detections in natural language alongside object detections. Use this capability to recognize complex conditions like hazards, anomalies, or operational states in real time. Situation custom insights deliver tailored, actionable insights without requiring coding or machine learning (ML) expertise.
For more information, see Custom insights.
November 2025
Real-time analysis (preview)
Azure AI Video Indexer enabled by Azure Arc introduces real-time analysis for video intelligence at the edge. Extract actionable insights from live video streams with low latency and complete data privacy while keeping all video data on-premises. Azure Arc provides centralized management and scalability.
For more information, see Live video analysis overview.
New capabilities in real-time analysis:
Agentic intelligence: A modular, agent-based architecture with specialized AI agents trained to detect specific events. Prebuilt agents support detections such as safety hazards, customer experience issues, or operational anomalies. Use the API to invoke an agent and provide instructions for the events you want to detect. You don't need to know which agent is best for the task. Azure AI Video Indexer routes queries to the appropriate agent.
Custom AI insights with natural language: Define custom detection logic by using natural language to monitor specific objects. You don't need technical expertise. Describe what you want to detect, and upload supporting images or negative examples to fine-tune detection. Available through the API and the Azure portal.
Event summarization: Generate end-of-shift summaries with your own focus to get a text summary of your video footage. This capability helps teams quickly review key events and operational outcomes.
Area of interest: Define specific regions within the video frame where you want to focus on the analysis.
For more information, see the following articles:
- Real-time video analysis overview
- Manage real-time analysis extensions
- Create custom insights
- Create an event summary for camera footage
- Mark an area of interest
March 2025
General availability of Azure AI Video Indexer enabled by Azure Arc
Azure AI Video Indexer enabled by Azure Arc is now generally available. You can run Azure AI Video Indexer in your edge environment to generate audio and video insights and textual summaries for videos in more than 30 source languages. For more information, see Azure AI Video Indexer enabled by Azure Arc overview.