2 hours ago
1
Google is turning its Gemini desktop app into something that feels less like a chatbot and more like an actual assistant. The latest addition, a voice dictation feature called “Speak to Window,” lets users bark commands at whatever app they happen to be working in, no tab-switching required.
The feature works by holding the Fn key. Hold the key, speak, and Gemini processes your voice command in the context of whatever application is currently in focus. Drafting an email, editing a document, comparing products: the idea is that Gemini becomes a layer on top of your entire desktop rather than a separate destination you have to visit.
How Speak to Window and Magic Pointer actually work
Speak to Window is rolling out alongside another feature called Magic Pointer. Where Speak to Window handles voice dictation across apps, Magic Pointer takes things a step further by combining screen pointing with contextual voice or text prompts.
You can literally point at something on your screen, ask Gemini about it, and get a response that understands what you’re looking at. Think image editing guidance, product comparisons, or pulling information from a chart without having to screenshot it and paste it into a chat window first.
Magic Pointer leverages Gemini’s vision-language capabilities. It was demonstrated in May 2026, showing how users could highlight elements on their display while simultaneously giving voice commands for context-aware assistance.
Testing reports for Speak to Window surfaced in June 2026, suggesting Google has been iterating on the feature for at least a few weeks before this broader announcement.
Why a desktop app matters for Google’s AI strategy
Google launched the Gemini macOS desktop app earlier in 2026. The entire premise was straightforward: give users on-demand AI assistance without forcing them to open a browser tab.
This is the same playbook that competitors have been running. OpenAI launched its own desktop app for ChatGPT, and Anthropic has been pushing Claude into similar territory.
The Fn key activation for Speak to Window is a deliberate design choice worth noting. It’s the same key that Apple uses for dictation on macOS, which means Google is essentially positioning Gemini as a replacement for Apple’s built-in dictation. Except instead of just transcribing speech to text, Gemini processes the intent behind what you’re saying and acts on it.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.
English (US) · 