wtf is qwen-mm-plugins?

Qwen-MM-Plugins is a set of add-ons that give AI coding agents the ability to understand images, video, audio, and other files, not just text. It plugs into several existing agent tools, including Claude Code, Codex, Qoder, OpenClaw, Qwen Code, and Gemini CLI, using a guided installer that adds each capability using that tool's normal install process. Each capability installs as a combination of a skill and, in most cases, a companion server the agent can call. The core capability lets an agent read images and video and visualize documents, code, data, and 3D files without needing an API key, though it does need ffmpeg installed for audio and video. Other capabilities add more: vision and OCR through Qwen's cloud models, web and reverse image search, building a searchable memory of long videos for question answering, generating and editing images, video, and audio, working inside Blender for 3D modeling and rendering, working inside FreeCAD for parametric CAD design, and creating narrated Chinese math and science explainer videos. Once a capability is installed, using it is meant to feel natural: you reference a file with an @ symbol in your message and describe what you want, such as summarizing a page of a PDF, transcribing a video with speaker labels, or identifying where a photo was taken, and the relevant tool is selected automatically. Some capabilities work entirely locally, while others need API keys from providers like DashScope, Serper, Exa, or Tavily, and some require desktop applications such as Blender or FreeCAD to already be installed on the system. The installer includes steps to configure credentials and verify that dependencies are in place. The project is released under the Apache 2.0 license, with separate attribution notices for the third-party Blender and FreeCAD integrations.