Ensuring SRS completeness is critical to preventing costly downstream errors and rework. We introduce an automated tool that ensembles three complementary LLMs—DeepSeek Chat, GPT-4o Mini, and Claude Sonnet 4—to detect and suggest remedies for missing requirements. The tool first generates a structured domain model from the SRS, then runs parallel external and internal completeness analysis using carefully crafted prompts. Users select which LLMs to invoke and choose among majority voting, weighted voting, or a Meta-LLM fusion to aggregate outputs. In experiments on four SRSs with seeded omissions, single models achieved only 0–52% recall, while our full ensemble consistently exceeded 75% (up to 100%) recall with 95–100% suggestion plausibility. These early findings highlight the potential of multi-LLM ensembles to dramatically outperform individual models and support next-generation requirements analysis tools through effective human-in-the-loop refinement.