← all repositories
Classic298/open-webui-plugins

Giving Open WebUI eyes, memory, and a mop

A curated plugin collection that makes Open WebUI remember reasoning chains, see images, and clean up its own database.

523 stars Python Other AI
open-webui-plugins
Velocity · 7d
+2.0
★ / day
Trend
accelerating
star history

What it does Eight self-contained plugins that slot into Open WebUI’s extension points—tools, filters, events, and skills—to add interactive visualizations, database pruning, email composition, and model capability bridges. Each lives in its own folder with a README explaining which components it needs and what problem it solves. The collection targets specific gaps in the default experience, from missing vision support to discarded reasoning context.

The interesting bit The Vision Bridge is particularly clever: it lets a text-only model work with images by stripping them to file-id markers in the request, then exposing an analyze_image tool that pings a separate vision model on demand. Meanwhile, Keep reasoning_content fixes a genuine platform bug where Open WebUI discards a reasoning model’s chain of thought before follow-up turns, causing mid-conversation errors.

Key highlights

  • Inline Visualizer v2 renders interactive HTML, SVG, and charts from Chart.js to Tone.js inline in chat, with a nine-color theme-aware palette and conversational drill-down.
  • Prune throttles background cleanup of old chats, orphaned uploads, and inactive users so a live instance never hiccups, with dry-run defaults and Redis-coordinated multi-replica support.
  • Vision Bridge gives text-only models on-demand image analysis by offloading vision work to a separate model without touching core code.
  • Keep reasoning_content feeds prior reasoning back into context for DeepSeek, Kimi, MiMo, and vLLM so tool calls don’t collapse mid-conversation.
  • MCP App Bridge renders MCP Apps (SEP-1865) as Rich UI embeds inline, injecting server-declared CSP headers without middleware changes.

Caveats

  • Several plugins require Open WebUI 0.10.0 or 0.11.0; version compatibility is explicitly noted per plugin but not summarized globally.
  • The Inline Visualizer has a legacy v1 alongside v2, which suggests the API surface may still be shifting.

Verdict Worth browsing if you run Open WebUI in production and have already hit its default limitations—especially around reasoning models, image handling, or database bloat. Skip it if the default Open WebUI already covers your use cases.

Frequently asked

What is Classic298/open-webui-plugins?
A curated plugin collection that makes Open WebUI remember reasoning chains, see images, and clean up its own database.
Is open-webui-plugins open source?
Yes — Classic298/open-webui-plugins is open source, released under the BSD-3-Clause license.
What language is open-webui-plugins written in?
Classic298/open-webui-plugins is primarily written in Python.
How popular is open-webui-plugins?
Classic298/open-webui-plugins has 523 stars on GitHub and is currently accelerating.
Where can I find open-webui-plugins?
Classic298/open-webui-plugins is on GitHub at https://github.com/Classic298/open-webui-plugins.

heatdrop uses Google Analytics to see which pages get read — nothing else. Your call. How we handle data.