Runtimes & apps Published:
Open WebUI 0.13 adds a tabbed workspace with terminal, browser and desktop, plus multi-model sub-agents

What happened
Open WebUI v0.13.0 turns the button at the top of a chat into a tabbed workspace for files, multiple terminal sessions, artifacts, previews, voice calls and comments, with a Git panel and, on a connected Open Terminal that supports them, Desktop and Browser tabs where you and the model share a computer-use or browser session. A model can now delegate work to other models chosen in the model editor, instead of only to copies of itself, and a single response can run up to 1024 rounds of tool calls, up from 256. Administrators get comment threads, a notification inbox, MCP Apps shown in a sandboxed frame with every tool call confirmed, and caps on knowledge base file counts and sizes. The release also contains security fixes not yet listed and database schema changes, so the project asks for a backup first and says multi-instance deployments must all be updated at once.
Why it matters locally: Self-hosted Open WebUI installs should update for the security fixes, but plan the upgrade: back up the database first and update every instance at the same time.
Three ways to read it
For the market
With terminals, a Git panel, a shared browser and desktop control beside the chat, Open WebUI is moving from a chat front end toward an agent workspace, the space hosted coding and computer-use assistants occupy. That likely makes it more attractive to teams that want those workflows on their own servers.
In plain words
Open WebUI is a free AI chat app many offices run on their own server. This version gives the AI a workbench next to the chat, with files, a browser and a desktop it can use while you watch, and lets one AI hand jobs to other AIs, like a team lead passing tasks to specialists.
For practitioners
Back up the database before upgrading and update all instances at the same time, since rolling updates are not supported for this schema change. Sub-agent limits now default to none (they were 20 concurrent, 20 in the background, 30 tool loops and 30000 characters of output), so set SUBAGENTS_MAX_* if local GPU capacity is tight, and CHAT_RESPONSE_MAX_TOOL_CALL_ITERATIONS=256 restores the old tool-call cap. KNOWLEDGE_MAX_FILE_COUNT, KNOWLEDGE_MAX_FILE_SIZE and KNOWLEDGE_MAX_UPLOAD_COUNT cap knowledge base growth.
Sprout's insight
Sprout compares this with other brands and with where real applications are heading. Free for members.
Sign in free to read it →Subscribe to Sprout's weekly
One email a week: the top stories and a taste of Sprout’s insights. Unsubscribe any time.
Source
Read the official source: Open WebUI v0.13.0 release notes (GitHub) ↗Summaries are written by OnPrem World from the official source. We link to the original and never republish it.
More news
- NVIDIA and Microsoft bring Nemotron 3.5 and RTX PRO 6000 Blackwell support to Azure Local
- Microsoft publishes Aion 1.0 Instruct, the 2.3B model that becomes Windows' built-in AI by end of January 2027
- RTX Spark compact desktops from ASUS, Dell and MSI go on sale with up to 128 GB of unified memory
- NVIDIA reports $115.8 billion quarter; highlights include a deal to acquire Hugging Face and a 64GB DGX Spark
- Open WebUI v0.13.4 moves personal connections into Tether, sandboxes agents on your own computer and runs local MCP servers