Vibeleaderboard
← All Intel
Intel / repo

open-webui v0.13.0

Source
github.com
Author
github-actions[bot]
Date
Why it matters

Open WebUI now supports computer-use workflows in a desktop tab and keeps terminals, files and artifacts alongside the chat. Self-hosters can try agentic desktop control with compatible models.

Key takeaways · AI-distilled
  • Open WebUI 0.13.0's Desktop tab shows and controls the desktop of a connected Open Terminal, so a model with desktop tools can run computer-use workflows while you watch and step in. It requires an Open Terminal connection with desktop support.
  • A new Browser tab in Chrome mode gives you and the model a shared browser session on a compatible Open Terminal, so the model can navigate sites, fill in forms and test the web apps it builds.
  • The release changes the database schema. The project recommends backing up first, and multi-worker or load-balanced deployments must update every instance at once, because rolling updates will cause failures.
  • Native now allows up to 1024 tool-call rounds per response, up from 256, and limits that previously defaulted to 20 concurrent and 30 tool loops each now default to no limit.
  • Apps now render in chat inside a frame that can reach only the sites the server declares, and every tool call the app wants to make waits for user confirmation.
Terms in this piece · Glossary
  • tool use — A model's ability to call external functions — run code, search the web, edit files — instead of only generating text.
  • subagent — A separate agent spawned by another to do one scoped piece of work in its own context, returning only the result.
  • MCP — The Model Context Protocol — an open standard that lets any AI assistant plug into any tool or data source without custom integration code.
  • sandbox — An isolated environment where AI-generated code or agent actions run without being able to touch anything real.
More from github-actions[bot]
Recommended reads
Comments

Checking sign-in…

Loading comments…