A curated collection of plugins for Open WebUI: tools, skills, filters, pipes, actions and events that extend your AI chat experience.
Each plugin lives in its own folder with a README explaining what it does, what components it includes, and how to set it up.
| Plugin | Description | Components | |
|---|---|---|---|
 |
Inline Visualizer v2 | 🔹 LIVE RENDERED 🔹 Your model draws interactive charts, dashboards, diagrams and mini-apps right in the chat, appearing live while the answer streams. Everything matches your light/dark theme automatically, and you can click a bar or a node and ask the model about it. Works with hand-written SVG/HTML and with Chart.js, D3, Vega-Lite, ECharts, Plotly, vis-network and Tone.js. | Tool + Skill |
 |
Prune | ⭐ NEW ⭐ Old chats, inactive users and orphaned files are cleaned out automatically in the background, so slowly that a live instance never notices. Dry-run by default, with an admin page at /prune to preview and run cleanups manually. Safe across replicas. |
Event |
 |
Interface Defaults | Open WebUI's Default Interface Settings page cannot push your defaults onto existing users. This adds that, a full factory reset, and defaults for the settings that page cannot reach: notifications, keyboard shortcuts, memory, the personal system prompt and the speech/voice block. | Event |
 |
Email Composer | Draft emails with your model and get a real email card in the chat: edit the rich text, adjust To/CC/BCC and priority, then download it as .eml or hand it to your mail app with one click. | Tool |
 |
MCP App Bridge | Use MCP Apps (SEP-1865) inside your chat: connect an MCP server and its tools bring their own interactive interface, rendered as an embedded panel right where the model called them. No middleware and no core changes needed. | Tool |
 |
Vision Bridge | Lets a text-only model work with images: attach a picture and the model asks a separate vision model to look at it on demand, then comes back with new questions about the same image any time. No core changes. | Filter + Tool |
| Inline Visualizer | 🗄️ LEGACY (v1) 🗄️ The original version: interactive charts and visualizations in chat, rendered once the answer finishes, with the same theme-aware design system and Chart.js/D3 support. | Tool + Skill | |
| Keep reasoning_content | Feeds a reasoning model its own prior chain of thought back, so DeepSeek / Kimi / MiMo / vLLM models stay coherent across tool calls and follow-up turns. Fixes the reasoning_content is missing break mid-tool-call. |
Filter |
| Type | What it does | Where to install |
|---|---|---|
| Tools | Give your model new capabilities it can call (web search, APIs, rendering) | Workspace → Tools |
| Skills | Structured instructions that teach a model how to do specified tasks or workflows | Workspace → Skills |
| Filters | Transform messages before they reach the model or before they're shown to you | Admin Panel → Functions |
| Pipes | Custom model endpoints. Proxy, merge, or create entirely new model behaviors | Admin Panel → Functions |
| Actions | Buttons that appear below messages for quick actions | Admin Panel → Functions |
| Events | React to system events as they fire (user signup, valve updates, lifecycle hooks). Can also register routes and serve standalone pages. | Admin Panel → Functions |
- Open the plugin's folder and read its README for specific instructions
- Each README lists the components (tool, skill, filter, etc.) and where to install them
- Some plugins are a single file, others are multi-component; the README will guide you
Each plugin folder is self-contained with all necessary files and documentation.
Found a bug or have an idea? Open an issue.