Open WebUI: Open WebUI v0.9.0 adds desktop app with task scheduling

For a long time, Open WebUI was primarily seen as a polished frontend for Ollama—useful, but tethered to the friction of Docker containers and terminal configs. The v0.9.0 release marks a strategic shift. By launching a native desktop application, they've removed the biggest barrier to entry for non-devs and casual power users. It’s no longer just a UI; it's a locally executable binary that can either stand alone or bridge to remote instances. Technically, the most impressive work is happening under the hood. The transition to a fully asynchronous backend across database and request paths addresses a common pain point in AI wrappers: the 'UI freeze' during heavy token generation or RAG indexing. The introduction of 'chat history memory culling'—essentially virtualizing the message list—shows a level of engineering maturity often ignored in the rush to add new LLM features. They are treating the interface like a professional IDE rather than a simple chat window. Product-wise, the addition of scheduled automations and a calendar workspace attempts to turn the LLM into an agentic assistant. Being able to schedule daily digests or periodic reports directly within the UI moves the tool from 'reactive' to 'proactive.' However, the reliance on internal scheduling requires careful resource management, which they've addressed with admin-level rate limits to prevent system overload. This is now a mandatory tool for anyone self-hosting LLMs who is tired of managing infrastructure just to get a clean interface. While the 'automation' features are still in their infancy compared to dedicated agent frameworks, the integration of a native app and professional-grade performance optimizations makes this one of the most practical AI interfaces available today.