ChatGPT’s voice mode is no longer a separate interface (techcrunch.com)

OpenAI has integrated ChatGPT’s voice mode directly into the main chat interface: instead of switching to a separate animated “blue circle” screen where you could only listen to responses, users can now speak and watch ChatGPT’s answers—including images, maps and earlier messages—appear in real time within the same conversation. The new voice experience is the default and is rolling out now across web and mobile apps; users who prefer the older isolated interface can re-enable “Separate mode” under Voice Mode in Settings. Note you still need to tap “end” to stop the voice session when you want to switch back to text. For the AI/ML community this is a meaningful UX and product improvement for multimodal conversational systems. Integrating audio input/output with live visual context reduces cognitive load, eases transitions between speech and text, and enables smoother workflows for tasks that combine spoken queries with images or maps. Practically, it improves accessibility and real-time inspection of model outputs (so you don’t miss or have to retrieve spoken replies), and it highlights design priorities for persistent context and multimodal rendering in deployed LLM interfaces.