Tech • AI • Robotics • Game
Gemini Has a New AI Super App | Here’s How to Master It
7/10TL;DR
Gemini has expanded its desktop app into a broader AI workspace by combining chat, autonomous tasks, coding, app integrations, and on-device shortcuts under Gemini 4 Argon.
KEY POINTS
Tasks now sit at the center of the app
Gemini now separates Chat from Task, making it easier to distinguish between asking questions and assigning work. The task mode is built around an always-on agent previously known as Spark, positioning the app as a tool that can execute actions rather than only generate replies.
Folder access turns Gemini into a local file agent
Users can connect specific folders on their computer and allow Gemini to read, create, and edit files inside them. That lets the agent work directly with business documents such as invoices, contracts, notes, and reports, then generate outputs like presentations or financial summaries inside the same folder.
Backups are built in for risky file operations
When a connected folder is used for a task, Gemini creates backups tied to that task. This adds a safeguard for users who allow the agent to modify local files, reducing the risk of accidental deletions or unwanted edits during automated work.
Remote task execution extends the desktop app to phones
Tasks can be started from a phone through the Gemini mobile app, including jobs that rely on files stored only on a computer. For local-file access, the computer must remain powered on with the desktop app open, and remote task execution must be enabled in settings.
Connectors broaden Gemini beyond Google services
A new Connectors section allows Gemini to work with third-party apps including Canva, Dropbox, Granola, and Wix. These integrations let the assistant pull information from external tools and, in some cases, create assets directly inside them, such as building a website within Wix.
Skills add reusable workflows
The app also includes Skills, which function as saved prompts or repeatable workflows. Users can create custom routines for recurring jobs such as brainstorming, email triage, or document handling, then trigger them quickly with a shortcut instead of rewriting detailed instructions each time.
Anti-Gravity brings coding into the main app
Anti-Gravity, Google’s coding environment for Gemini, is now integrated directly into the desktop app instead of remaining separate. The coding tool can generate software projects from plain-language prompts and handle setup tasks on the computer, lowering the barrier for non-programmers who want to build games, apps, or utilities.
Gemini 4 Argon powers the upgrade
The new experience runs on Gemini 4 Argon, described as Google’s most powerful current model. It is positioned as especially strong for coding, writing, research, and document-heavy knowledge work, though heavier use may consume paid-plan limits faster than lighter models such as Flash.
Speak to Window adds system-level voice and screen control
A feature called Speak to Window lets users invoke Gemini with keyboard shortcuts, dictate text into active windows, or send the contents of a selected window into chat for follow-up work. That enables quick actions such as turning screenshots into idea lists, generating titles, or inserting dictated notes without manually copying material between apps.
The app is broader, but still incomplete
Beyond core chat, tasking, and coding, Gemini also supports image generation, scheduled tasks, daily briefs, notebooks, music, and video creation. Still, notable gaps remain, including the absence of a desktop live mode, limited native browser capability, and a need for more third-party connectors.
CONCLUSION
The latest Gemini updates push the app beyond a standard chatbot into a unified assistant for research, automation, coding, and file-based work. Its usefulness now depends less on conversation alone and more on how effectively users connect it to their documents, devices, and external tools.
Ask a question