Open WebUI 0.13.0 brings files, terminals, browser sessions and desktop control into a tabbed workspace alongside the chat. For people building with AI, the shift is from asking a model for help to keeping more of the work, and the tools around it, in one place.
Open WebUI Watch analysis
What happened
The project’s 0.13.0 release notes say the version was released on 3 November. Users can open tabs for terminals, files, browser sessions and supported desktops, then switch between them or expand the workspace on larger screens.
The desktop and browser features depend on a connected Open Terminal with the relevant support enabled. In a shared browser session, people can watch and interact with the page while a model uses browser tools. The desktop tab can show and control a connected desktop, including mouse and keyboard input.
Our top picks
- A workspace that keeps chat and work together
Tabs can hold files, terminal sessions, previews, voice calls and comments beside the conversation. - A shared browser for people and models
On a compatible connection, a person can watch or take over while a model navigates websites and tests web apps. - A desktop tab for computer use
Users can view and control a connected desktop, with model-driven workflows available when the required tools are enabled. - Git tools in the side panel
In a repository, users can review changes, stage or discard them, manage branches and worktrees, and work with pull requests. - Side chats that leave the main thread tidy
A temporary conversation can use the chat so far as context without adding the detour to the main conversation. - Comments and an inbox
Comments, replies and mentions can appear across workspace items, while the inbox collects notifications and other updates. - Interactive MCP apps in chat
Apps from MCP servers can appear in a sandboxed frame, and each tool call they request waits for user confirmation. - Previews for 3D files
Supported GLB, glTF, STL, OBJ and PLY files can be opened in a viewer that users can rotate and zoom.
Why it matters
This release puts model-assisted work closer to the tools where it happens: code, web apps, files and desktops. That could save the familiar shuffle between chat windows and workspaces, while giving people a way to watch and intervene in some computer-use workflows.
The practical catch is that several headline features rely on a connected Open Terminal and enabled support or permissions. The notes describe what shipped, not how reliably every setup will handle it.
Our read
This is a meaningful step towards an AI workspace rather than a chat window with accessories. The browser, desktop and Git features are the standouts because they connect conversation to work people can inspect. If you use Open WebUI, check the release notes against your setup before expecting every tab to work out of the box. The model may have a desk now; it still needs the right keys.
What to watch
- Which Open Terminal setups support the new desktop and browser workflows.
- How users and administrators manage permissions for computer-use features and comments.
- Whether the Git panel and interactive MCP apps fit cleanly into everyday workflows.
Discussion spark: Would you trust an AI workspace to work across your browser and desktop if you could watch and take over, or should computer-use tools stay separate from chat?
Sources and evidence
- Releases · open-webui/open-webui · GitHub (Publication date not supplied)
not affiliated with, endorsed by, or speaking for Open WebUI or its contributors