Open WebUI v0.13.0 Released: Tabbed Workspaces & Security Fixes

Open WebUI v0.13.0 Released: Tabbed Workspaces & Security Fixes

At a Glance

Item Value
Repository open-webui/open-webui
Version v0.13.0
Published 2026-11-03
License NOASSERTION
Source type Primary source (the publisher itself)

Values determined by this site’s code when the information was collected. Dates are JST.

Overview

open-webui/open-webui is a user-friendly AI interface that supports various backends such as Ollama and the OpenAI API.

In this v0.13.0 release, a “Tabbed Workspace" has been introduced to the chat screen, allowing files, terminals, browsers, and desktop operations to be handled integrally within a single screen. This makes it possible to seamlessly execute generated code and test websites while conducting a dialogue with the AI.

The most critical change is that security fixes are included. This release includes fixes for undisclosed security vulnerabilities. All users are strongly encouraged to update as soon as possible to ensure safety.

Key Changes

Introduction of Tabbed Workspaces

From the “Tabs" button at the top of the chat, you can now open a workspace where you can manage files, multiple terminal sessions, artifacts, previews, voice calls, comments, and more. Tabs can be added from the “+" menu or reordered by dragging. When doing coding, data analysis, or reviewing generated documents, you can manage the contents of the conversation and their working files in one place, greatly improving work efficiency.

Enhanced Computer Use and Browser Access

The “Open Terminal Desktop" tab allows you to view and control the desktop provided by a connected Open Terminal. Support for mouse, keyboard, clipboard, and audio where available is included. When using a model that can call desktop tools, you can track the computer operation workflow performed by the model directly on the screen or have it handle manual inputs on your behalf. This requires an Open Terminal connection with desktop support enabled.

Additionally, the “Browser" tab allows you to directly open websites or local development servers alongside the chat. Featuring navigation controls, multiple tabs, and a mobile viewport, Chrome mode allows for browser automation such as site navigation, form interaction, and testing web apps built by the model through a shared browser session with the model.

Security Fixes (Important)

This release includes security-related fixes. These fixes are scheduled to be disclosed publicly after CVEs (Common Vulnerabilities and Exposures) are assigned, but a prompt update is strongly recommended for all deployments to apply the latest patches.

Evolution of Multi-Agent Workflows

In the model editor, you can now select another model as a “sub-agent" to which the parent model delegates tasks. This allows you to build advanced multi-agent workflows where different models are assigned to specific roles, such as “Model A for research, Model B for coding, and Model C for review." In previous versions, sub-agents always used the same model as the parent model, but this update significantly improves flexibility.

In-Chat Display of MCP Apps

When running tools accompanied by interactive apps provided by MCP (Model Context Protocol) servers, the app is now displayed directly within the chat as a sandboxed frame. All tool calls that the app attempts to execute await user confirmation. This allows you to safely operate external tool interfaces while maintaining the context of the chat.

Git Integration and Real-Time Terminal Updates

A Git panel has been added to the side panel. If the selected terminal folder is a Git repository, you can check changes (diffs), stage, commit, manage branches and worktrees, push and pull, view history, and even perform GitHub pull request operations. In addition, in supported terminals, the Git panel, running processes, and list of open ports now update instantly the moment a change occurs rather than updating every few seconds.

Comment Feature and Notification Inbox

When administrators enable settings (Admin Settings > General or the ENABLE_COMMENTS environment variable), users can leave comments, replies, reactions, and mentions on chats, notes, models, knowledge bases, prompts, skills, and tools. A new “Inbox" has also been added to the sidebar to centrally manage notifications such as comment replies, mentions, channel interactions, automation failures, and upcoming calendar events.

Enhanced Administrator Features

In knowledge base management, administrators can now set limits on the number of files, individual file sizes, and the number of files in a single upload. A feature has also been added allowing models to create private Python tools on behalf of users, though this is restricted to administrators and users permitted to create tools.

Hardware Requirements

In Open WebUI Desktop 0.1.x (independently installed apps for macOS, Windows, and Linux), improvements have been made regarding the setup and updating of local models using NVIDIA and AMD GPUs. This makes running local AI workflows smoother.

How to Get It

If you are using a Python environment, you can install or update with the following commands:

pip install open-webui
open-webui serve

If you are using Docker, you can run it with the following command:

docker run -d -p 3000:8080 --add-host=host.docker.internal:host-gateway -v open-webui:/app/backend/data --name open-webui --restart always ghcr.io/open-webui/open-webui:main

If you are using the desktop app, please download the latest version from the GitHub repository.

Related Articles

What to Read Next

Sources