ComfyUI_Qwen3-VL-Instruct

Why we included this project

ComfyUI users who want vision-language capabilities inside their node graph can install this custom node and get the Qwen3-VL-Instruct models working directly in their workflows. It takes text, single images, multiple images, or video as input, and returns captions or responses, so visual understanding can feed into a larger pipeline instead of living in a separate tool. The bundled example workflows cover all four query modes, and models download automatically on first run, so you don't have to fetch them by hand. Teams that need image captioning or video summarization inside ComfyUI will find this a practical addition. One thing to plan for: you may need the companion Display Text node from the author's MiniCPM-V addon to see outputs.