ChatGPT Projects get a major upgrade: Smart workspaces, deep research, and voice mode now available

ChatGPT

OpenAI has announced a powerful new upgrade to ChatGPT Projects, transforming them into smarter, more collaborative workspaces that streamline long-term efforts like writing, planning, and research.

According to a blog post by OpenAI, users can now harness tools like Deep Research, which blends chats, files, and instructions with live web sources to produce in-depth, context-rich results. It’s a major leap for anyone using ChatGPT for ongoing work, especially researchers and professionals managing evolving tasks, said the company.

Voice Mode is also rolling out inside projects, allowing hands-free interaction or spoken questions about uploaded files.

Sharing gets easier too, as per OpenAI. Users can now create a unique URL for any individual chat in a project, making it simple to collaborate or showcase results. Importantly, these shared links only expose that specific chat, keeping all other project elements private.

On the mobile front, the ChatGPT app is getting some much-needed love: users can now upload files and switch models directly within the app. Do keep in mind that you will need the latest update.

One of the most key changes is improved memory within projects. For Pro and Plus users, ChatGPT can now remember past chats and reference them to stay on-topic—ideal for multi-week or multi-person workflows. Unlimited projects are supported, each allowing up to 20 files.

How to create a project in ChatGPT?

Creating a project is easy: just hit “New Project” in the sidebar and give it a name. Existing chats can also be dragged into a project to inherit its context, memory, and files. Users can upload PDFs, spreadsheets, or images, and even set role-based instructions (like “Act as my startup advisor”).

With tools like image generation, code drafting via Canvas, and a unified research interface, Projects now look less like folders and more like a true AI workspace.