Titikey
HomeTips & TricksChatGPTChatGPT New Features Overview: macOS Desktop App and Direct Uploads from Google Drive/OneDrive

ChatGPT New Features Overview: macOS Desktop App and Direct Uploads from Google Drive/OneDrive

3/20/2026
ChatGPT

Recent ChatGPT updates lean more toward real-world usability—not just stronger models, but tighter integration with files, voice, and desktop workflows. Below is a practical, use-case-based roundup of the ChatGPT features and changes most worth paying attention to.

Mac desktop app: Put ChatGPT on a keyboard shortcut

ChatGPT now has a macOS app. A common way to open it is pressing Option + Space to bring it up instantly—no need to switch browser tabs. It supports uploading files and photos directly from your desktop, and you can also have voice conversations inside the app, which works well when you’re reviewing materials while asking questions, or iterating and confirming edits as you go.

Another useful improvement is smoother chat history search: if you use ChatGPT as a “work log,” it feels more like a native tool than the web version when you need to pull up a previous plan, code snippet, or writing draft.

Direct cloud drive uploads: Feed Google Drive and OneDrive files straight into ChatGPT

For data-analysis tasks, ChatGPT has added the ability to upload files directly from Google Drive and Microsoft OneDrive—saving you the back-and-forth of downloading locally and re-uploading. For spreadsheets and chart-heavy content, you can keep asking follow-up questions in the same conversation, have ChatGPT summarize and compare results, and export customized charts for presentations.

If you often use ChatGPT for weekly reports, ops metrics, or budget sheets, this kind of “cloud direct upload” makes it feel less like simple chat and more like a lightweight data workbench.

More natural voice: Advanced Voice Mode rolling out

ChatGPT is rolling out more lifelike voice conversations (Advanced Voice Mode). The key difference is that responses sound closer to real human pacing and emotion, and it can work with vision features for multimodal interaction. This feature is currently being released gradually—typically first to a subset of subscribers, then expanding over time.

At the same time, OpenAI is being more cautious about voice compliance: after a dispute over the “Sky” voice sounding similar, that voice was temporarily paused—showing that as voice gets better, the platform is also increasing controllability and safety checks.

No-login access and SearchGPT: ChatGPT moves closer to a “search entry point”

ChatGPT can now be used without an account under certain conditions, but the experience varies—for example, you may not be able to save or share chats, or use features that require an account. For new users, this lowers the barrier to trying it; for power users, logging in is still better for full functionality.

Separately, OpenAI is testing a SearchGPT prototype that brings ChatGPT-style Q&A into a search scenario, aiming to provide more timely answers from across the web and support follow-up questions. It’s currently limited to a small user group and select publisher partners to gather feedback and calibrate the experience.