We demonstrate a real-time, diffusion-based speech enhancement system capable of removing background noise in a reverberant environment using a single microphone. Building on recent advances in diffusion models for speech enhancement, the proposed system achieves sub-second latency while preserving speech quality in challenging conference scenarios, e.g. loud babble noise. Attendees will personally experience the difference between turning on/off the speech enhancement system and regulating the quality/latency trade off. This demonstration showcases, for the first time, an online, diffusion-based enhancement framework suitable for live events.