Skip to content

Classic298/open-webui-plugins

Repository files navigation

🧩 Open WebUI Plugins

A curated collection of plugins for Open WebUI — tools, skills, filters, pipes, and actions 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.


Plugins

Plugin Description Components
Inline Visualizer v2 🔹 LIVE RENDERED 🔹 interactive HTML/SVG visualizations inline in chat. Full design system with theme-aware colors, SVG utilities, pre-styled bare HTML (forms, tables, <details>, <dl>, …), a 9-color data-accent palette, and six interactive bridges for conversational drill-down - talk to your visualizations, dashboards, charts, maps etc. - First-class support for Chart.js, D3, Vega-Lite, ECharts, Plotly, vis-network, and Tone.js. Tool + Skill
Keep reasoning_content NEW ⭐ Keeps your reasoning model's chain of thought intact across tool calls and follow-up turns, so it no longer "forgets" why it called a tool a moment ago or breaks mid-tool-call with a reasoning_content is missing error. Open WebUI normally throws away the model's prior reasoning before sending the next request; this filter feeds it back, so DeepSeek / Kimi / MiMo / vLLM and other reasoning models stay coherent across an entire conversation. Filter
Email Composer AI-powered email drafting with an interactive Rich UI card. Rich text editing, To/CC/BCC chips, priority, download .eml, one-click send via mailto. Tool
MCP App Bridge Renders MCP Apps (SEP-1865) as Rich UI embeds. Connects to MCP servers, calls tools with ui:// resources, injects server-declared CSP, and renders the HTML inline — no middleware changes needed. Tool
Inline Visualizer 🗄️ LEGACY (v1) 🗄️ Interactive HTML/SVG visualizations inline in chat. Full design system with theme-aware colors, SVG utilities, Chart.js/D3 support, and a sendPrompt bridge for conversational drill-down. Tool + Skill

Plugin Types

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

How to Install

  1. Open the plugin's folder and read its README for specific instructions
  2. Each README lists the components (tool, skill, filter, etc.) and where to install them
  3. 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.


Contributing

Found a bug or have an idea? Open an issue.