Qwen-MM-Plugins

English · 中文

Native multimodal plugins for Qwen models. Make any agent harness multimodal-native.

Architecture

Install

The guided installer supports Claude Code, CodeBuddy, Codex, Qoder, OpenClaw, Qwen Code, and Gemini CLI. Shared configuration lives in ~/.qwen-mm-plugins/config.

In-app setup for WorkBuddy, QoderWork, and QwenWork, plus manual setup for DeepSeek Harness, Hermes Agent, opencode, pi, and QwenPaw, is documented in the other harness guide.

curl -fsSL https://raw.githubusercontent.com/QwenLM/Qwen-MM-Plugins/main/install.sh | bash

Update the capabilities already installed in one harness:

curl -fsSL https://raw.githubusercontent.com/QwenLM/Qwen-MM-Plugins/main/install.sh | bash -s -- update

Released capabilities use independent, immutable tags. For local checkout installs, rollback, manual skill + MCP setup, dependencies, and Windows/WSL2, see the installation guide.

DeepSeek Harness (DSH)

The source repository includes a DSH carrier at packages/dsh. It mounts selected Qwen capabilities through the official @deepseek-ai/dsh-mcp-client; the Python MCP servers still come from this repository's uvx package source. Install the carrier in the target profile:

dsh plugin --profile web add @omdsh-dev/qwen-mm-plugins-dsh

The carrier requires the native-media DSH client line and relies on its built-in attachment admission and finalizeContent support; it does not carry a compatibility layer for older MCP clients. See the DSH setup guide for profile configuration and source-checkout mode.

Capabilities

Each capability is installed independently as a Skill plus an optional MCP server. Its install name is qwen-mm-plugins-<capability>.

Capability	Use case	Main requirements	Cookbook
core	Read images and video; visualize documents, code, data, 3D files, and more	No API key; ffmpeg for audio/video; format-specific apps as needed	Cookbook
api	Qwen VL/Omni vision, OCR, grounding, ASR, segmentation, and audio-video understanding	DashScope; ffmpeg for local audio/video	Cookbook
search	Web search, page extraction, and reverse-image search	Serper, Exa, or Tavily key; image search requires Serper	Cookbook
video-memory	Build hierarchical memory for long-video QA	DashScope; ffmpeg/ffprobe for builds	Cookbook
video-edit	Image, video, and audio generation with editing workflows	DashScope; ffmpeg + Node/Chromium for full edits	Cookbook
blender	Model, texture, light, and render in Blender	Blender; Xvfb on headless Linux	Cookbook
freecad	Parametric CAD, STEP/STL, and FEM workflows	FreeCAD; CalculiX for FEM; Xvfb on headless Linux	Cookbook
edu-agent	Create Chinese math/science explainer videos and interactive pages	Skill-only; Node/Chromium + ffmpeg; DashScope for narrated video	Cookbook

Try it

After installing a capability, reference a file and ask naturally; the Skill selects the relevant MCP tool.

@report.pdf          Summarize page 3 and extract its table.
@meeting.mp4         Transcribe this with speaker labels and timestamps.
@place.jpg           Identify where this photo was taken and verify it on the web.
@lecture-2h.mp4      List the main points with timestamps.

core reads media at dynamic resolution, so manual resizing is normally unnecessary.

Requirements and configuration

• uv provides uvx, which installs Python dependencies on demand.
• Local core tools need no API key in the default native-image mode. Text-only caption fallback, cloud, and search capabilities need their provider credentials.
• Video, document, browser, Blender, and FreeCAD workflows may need system applications.

Run the installer's Configure and Verify actions to set credentials and check dependencies. See Installation for prerequisites and the configuration reference for every setting.

Documentation

• Installation
• Configuration
• Contributing · Local development
• Add a capability · Testing

License

Apache-2.0 — see LICENSE. Third-party attribution for the Blender and FreeCAD integrations is recorded in their respective Blender and FreeCAD notices.