Recent advancements in Large language models (LLMs) have demonstrated remarkable capabilities across diverse domains. While they exhibit strong zero-shot performance on various tasks, LLMs' effectiveness in music-related applications remains limited due to the relatively small proportion of music-specific knowledge in their training data. To address this limitation, we propose MusT-RAG, a comprehensive framework based on Retrieval Augmented Generation (RAG) to adapt general-purpose LLMs for text-only music question answering (MQA) tasks. RAG is a technique that provides external knowledge to LResearch goal: How does the multimodal capture component in Expert Mind affect VQA accuracy on domain-specific datasets compared to text-only RAG baselines?Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 9.2/10.