Existing knowledge graphs often impose substantial modification constraints due to stringent accuracy requirements, and there are few multimodal graphs in Traditional Chinese Medicine (TCM), hindering timely incorporation of cutting-edge domain knowledge. To resolve these challenges, we propose QiboGraph, a multimodal TCM knowledge graph featuring crowdsourced collaborative modification. We also introduce a two-stage alignment pipeline assigning images initial weights, to accelerate the convergency of the graph. Finally, a practical system is developed to demonstrate two typical use cases of QiboGraph.