ChatGPT Projects gets voice mode, Deep Research, and mobile upgrades

Projects just got a lot more useful for serious work. OpenAI announced on X that it is rolling out several new capabilities to projects in ChatGPT to help users stay focused on complex tasks.

Deep research arrives, mobile catches up

This update rounds out the feature set across both desktop and mobile.

First, projects now support Deep research, letting you run comprehensive investigations directly within your project's context. Voice mode is also enabled, so you can talk through your project materials hands-free.

Memory inside projects has been upgraded as well. ChatGPT can now reference past chats from the same project, cutting down the need to constantly re-explain background details.

On mobile, OpenAI added the ability to upload files directly into projects and access the model selector, finally bringing the app closer to feature parity with the web.

Why it matters

ChatGPT projects are steadily evolving from static organizational folders into genuine interactive workspaces.

We think adding Deep research and shared project memory solves one of the biggest friction points for power users: context loss across long-running tasks. Giving mobile users the ability to switch models and drop files into projects on the fly also ensures that your workflow stays intact when stepping away from the desk.

This move puts even more pressure on standalone productivity and workspace tools. However, the real test will be execution: as projects pile up with weeks of files and messy conversations, OpenAI will need to prove its memory retrieval stays accurate without hallucinating old context.