> ## Documentation Index
> Fetch the complete documentation index at: https://docs.worlds.io/llms.txt
> Use this file to discover all available pages before exploring further.

# 3.3.1

May 15, 2026

***

#### **Dashboards**

Several new enhancements enable deeper insights from user-created dashboards and configured visualizations:

* **Stacked Widgets** — Data visualizations can now be grouped in configuration, for faster and easier setup and improved visual clarity in dashboard layouts.
* **Time Groupings** — Dashboards now support 30-minute breakpoints for custom time groupings, giving teams more granular control over shift changes and other categorical time comparisons.
* **Complete Tooltips on Hover** — Bar and column widgets now display all fields in the hover tooltip rather than a sample, making it easier to interpret complex data without leaving the dashboard.
* **Heatmap Enhancements** — Event durations can now be visualized in the heatmap widget, for when duration of activity drives more insight than volume.
* **Time Formatting** — Users can now choose from additional display format options for time fields across dashboard widgets.

***

#### **Agentic Analysis**

We continued expanding our agentic capabilities with improvements to transparency, discoverability, and interactivity:

* **Real-Time Agent Progress** — AI now surfaces live progress and skill results as analysis runs, giving users visibility into what the agent is doing and why.
* **New Conversation Button** — When you're ready to stop digging deeper and start fresh on a new inquiry, easily begin new AI sessions with fresh context.

***

#### **Storage Management**

We expanded storage configuration to better support multi-site deployments:

* **Per-Site Storage** — Storage credentials can now be configured and managed independently per site, giving administrators more flexibility in how footage is stored across locations.

***

#### **Usability**

* **Personal Access Tokens** — Users can now create and manage personal access tokens that match their permissions, enabling self-serve API authentication without requiring an admin.
* **Video Upload Limit** — The frontend video upload limit has been increased to 5 GiB, aligning with backend capacity and supporting larger file uploads.

***

#### **Bug Fixes & Performance Improvements**

The following issues have been resolved in this release:

* **Inference Scalability** — Improved memory usage and startup times, enabling detectors to share GPU resources for more optimal performance and cost.
* **Tracking Performance** — Tracker now performs significantly better for stationary objects and long-running detections, allowing for more stable and reliable downstream workflows.
* **Tagging UI Performance** — Resolved a significant performance degradation affecting tag and video load times in the Tagging UI.
* **Camera Editing** — Resolved a "Resource Not Found" error that occurred when editing a camera from Builder.
* **Change Latency** — Fixed issues in tagging, indexing, and agent management that could cause committed changes to not display until a page refresh.
* **Video Timestamps** — Fixed an issue where uploaded images rendered with an incorrect Unix epoch 0 timestamp.