Open WebUI 0.9.0: Free Desktop App, No Docker Needed

Open WebUI 0.9.0, the free local AI desktop app, just graduated from a browser tab to a full desktop application — and the upgrade costs exactly $0. Version 0.9.0 ships a native app for Mac, Windows, and Linux with zero telemetry (meaning no conversation data is sent to any server), no Docker required, and built-in AI automation features that compete directly with ChatGPT plugins and Microsoft Copilot.

This isn't a cosmetic release. The update adds a calendar workspace, scheduled AI automations, system-wide voice input, and a floating chat overlay that works across your entire desktop — not just inside a browser window.

Open WebUI Desktop App: No Docker, Zero Setup Required

For years, running Open WebUI meant wrestling with Docker (a container system that packages apps so they run identically on any machine) or asking someone on your team to set it up. Version 0.9.0 eliminates that barrier entirely.

The new desktop app installs like any normal program on macOS, Windows, or Linux. Once it's running:

• Offline mode activates automatically after your first launch — no continuous internet connection required
• Automatic updates run in the background, so you never manually download a new release
• Zero telemetry — none of your conversation data leaves your machine or gets logged on a remote server
• Multi-server switching — connect to multiple AI backends (a local Ollama instance, a remote Open WebUI server, Azure OpenAI) and swap between them instantly from the sidebar

The app also introduces a system-wide floating chat bar — a compact overlay you can summon on top of any application. Press Shift+Cmd+I on Mac or Shift+Ctrl+I on Windows and Linux. Push-to-talk voice input works directly from this bar, so you can dictate a question without switching to a chat window at all. A screen-awake feature keeps the display on during long transcriptions so dictation doesn't cut off mid-sentence.

Calendar, AI Automations, and Task Tracking — Built Into Your Chat

The features that push Open WebUI 0.9.0 past "just a chat interface" are the productivity tools woven directly around the conversation view. You don't need a separate app for any of them.

Scheduled AI Automations

Scheduled chat automations let AI models run recurring tasks on a timer — daily news digests, weekly summary reports, periodic data checks — without you opening a new conversation each time. The automation runs, delivers its output, and logs it, entirely on schedule. You can create, update, list, pause, and delete automations using plain language inside a conversation.

Administrators can cap how many automations each user creates with the AUTOMATION_MAX_COUNT environment variable and enforce minimum time between runs using AUTOMATION_MIN_INTERVAL — useful for teams sharing a single server instance.

Full Calendar Workspace

A complete calendar view now lives inside Open WebUI. Create events, set recurring schedules, and receive reminders up to 1 hour before an event starts. Alerts fire through three simultaneous channels: in-app toasts (small pop-up notifications at the screen edge), browser notifications, and outbound webhooks (automated pings to external services like Slack or a custom script). Administrators tune the reminder engine using SCHEDULER_POLL_INTERVAL (how often the system checks for upcoming events) and CALENDAR_ALERT_LOOKAHEAD_MINUTES (how far ahead to scan).

Task Management in Conversations

AI models can now create, update, and track tasks directly inside a conversation thread. Ask your model to capture a follow-up item and it adds it to a persistent task list accessible from the sidebar. No copy-pasting into a separate app, no tab switching — the task lives alongside the conversation that created it.

Everything New in Open WebUI 0.9.0

Beyond the flagship features, the release packs a dense set of quality-of-life improvements:

• Mistral TTS (Text-to-Speech — the technology that converts written responses into spoken audio) is now an admin-configurable provider option alongside existing voice integrations
• Azure OpenAI connections now support the newer /openai/v1 endpoint format, fixing compatibility with updated Azure-hosted deployments
• Ollama proxy (Ollama is a free tool that runs large language models locally on your computer) now supports the Responses API via /v1/responses
• File reuse tab: previously uploaded files are accessible from a new Files tab in the chat input — no re-uploading the same document every session
• Emoji shortcodes: typing a colon triggers an emoji suggestion menu, with recently used emojis pinned at the top of the picker
• Mobile swipe-to-reply: swipe right on any message to reply to it directly
• Async backend: the core server now handles requests asynchronously (processing multiple conversations at the same time instead of in a queue), significantly improving response speed under concurrent load
• Pinned notes: notes can be pinned to the sidebar for quick access across sessions without opening a separate document
• WebSocket reconnection feedback: when your connection drops and reconnects (WebSocket is the live communication channel between the browser and server), the app now shows visible status so you know the session recovered
• Chat history memory culling: off-screen messages in very long conversations are automatically unloaded from memory and reloaded when you scroll back, keeping performance stable in hours-long sessions

One Critical Warning Before You Upgrade Open WebUI

Version 0.9.0 includes database schema changes — the underlying structure of how Open WebUI stores data has been updated. If you're running an existing installation, especially a shared or production deployment (one that serves real users), back up your database before upgrading. Do not skip this step.

Rolling updates (upgrading one server at a time while others stay live) are not supported in this release. In multi-server or load-balanced setups (where several machines share the work), all instances must update simultaneously. Updating only some of them causes application failures from schema incompatibility between old and new versions. Plan a maintenance window if your team shares a single Open WebUI instance.

For first-time desktop app users, none of this applies — you're starting fresh with a clean database and no migration needed.

Download Open WebUI Free — No Terminal, No Subscription

If Docker has been keeping you from trying Open WebUI, 0.9.0 removes that excuse completely. Download the desktop app from github.com/open-webui/desktop and install it like any other program. Connect it to a local AI model running through Ollama on your machine, point it at a remote server, or use the multi-server sidebar to switch between both. Your first scheduled AI automation — a daily digest, a recurring report — can be running within minutes of install. Zero telemetry, zero subscription, zero Docker commands required.

Related Content — Get Started with Local AI | AI Automation Guides | More News