Dictate to Gemini

Can you talk to Gemini already?

Yes. Gemini has voice input built in. In the Gemini app on your phone you tap the microphone and speak your question; on a supported device you can even start hands-free. On the web at gemini.google.com there's a microphone in the prompt area too. So if your goal is "ask Gemini something out loud and get an answer," you don't need anything else. The feature ships in the product.

I want to be clear about that up front, because a lot of articles for this keyword skip it and go straight to selling you something. Gemini's own voice input is genuinely fine for quick spoken questions. "What's the weather in Brasov," "convert 200 grams to cups," "summarise this article" — talk, listen, done. If that's the whole of what you do, this is a short article: use the mic that's already there.

The gap shows up the moment your prompt gets longer or more deliberate, and the moment you want to use the exact same habit somewhere that isn't Gemini. That's the rest of this piece.

Why type the prompt instead of just talking

When you talk straight into a chatbot's microphone, your words go in and the answer comes out — but you never see the prompt as editable text first. For a one-line question that's fine. For a careful prompt it isn't. You said "summarise this in a formal tone," the mic heard "summarise this in a formal town," and now the model is doing something slightly wrong and you're re-recording the whole thing. Typing the prompt into the box first means you read it before it runs.

The other reason is that the habit travels. A hotkey that types your speech into a text box doesn't care which box. Gemini today, ChatGPT tomorrow, a Claude prompt the day after, a Slack message, a commit description, an email to your kid's teacher. One muscle memory, every app. Gemini's mic only works inside Gemini, which is exactly what you'd expect, but it means a second tool that works everywhere isn't redundant — it's a different shape.

The fastest way: a system-wide hotkey

Here's where a dedicated app changes things. Whisper by Remskill is not a Gemini plugin, a Chrome extension, or a Google integration. It's a desktop app that behaves like your keyboard: press a hotkey, speak, and the transcript is pasted at your cursor — in any app, the Gemini prompt box included. It works the same in the browser and on the desktop, because as far as your computer is concerned, you're just typing.

Setup is short:

That's the whole loop. The transcript appears in the box, you read it, you fix anything off, you send.

Dictate a prompt straight into Gemini

Once it's running, the experience is unremarkable in the best way. You put your cursor in Gemini's prompt box, hold the key, say your prompt, let go. A second or so later the text is sitting there as if you'd typed it. Now you do the thing the live mic doesn't let you do: you read it. Long prompts especially — the three paragraphs of background plus the actual question — are far easier to get right when you can see them before they run.

Because the local transcription runs on your machine (pure-Rust, no Python sidecar, no server in the loop), it works offline, and the audio of whatever you're dictating never leaves your laptop. You're typing the prompt with your voice and then sending it to Gemini yourself — the dictation step and the AI step are separate, which is the whole reason you get to edit in between.

AI cleanup strips the "ums"

Spoken language is messy. You say "um," you restart sentences, you trail off into a noise that means "you know what I mean." Whisper has an optional AI enhancement step that trims filler and tidies the phrasing before it pastes. So "uh, can you, can you write me like a, a short email declining this" becomes "Can you write me a short email declining this." A cleaner prompt tends to get a cleaner answer, so this one actually matters more for AI prompting than for plain notes.

That cleanup runs locally through Ollama, free, on your own machine. Pro users can route it through the cloud instead, but the filler-cleanup benefit doesn't need Pro; it's there in the free local pipeline. You can also turn it off and paste the raw transcript, which is fine when you're going to read and edit the prompt anyway.

The same hotkey works in ChatGPT, Claude, anywhere

This is the part that makes the separate tool worth it. The hotkey doesn't know what Gemini is. It pastes text wherever your cursor sits, so the exact habit you just learned works in ChatGPT, in Claude, in Perplexity, in a Google Doc, in your email client, in the comment box of a pull request. You learn one move and stop thinking about which AI you're prompting today. Gemini's built-in mic, by design, only helps inside Gemini.

I bounce between three or four AI tools in a normal week depending on what I'm doing. The thing I did not want was a different voice button in each one, each behaving slightly differently. One hotkey that types into all of them is less to remember, and it keeps working when one of them ships a redesign.

What the full Whisper app looks like

The hotkey is the part you'll use most, but there's a settings surface behind it. You pick your transcription engine: Whisper models, whose multilingual variants cover 99 languages and the English-only .en builds cover exactly one, or NVIDIA's Parakeet, about 600 MB, 5 to 10 times faster than Whisper on CPU, covering English plus 24 European languages. You set your hotkey, manage history, and save presets. None of that is required to dictate one prompt into Gemini. It's there when you want to tune.

When Gemini's own mic is enough

If you only want to ask Gemini quick spoken questions, skip the separate app entirely. Open Gemini, tap or click the microphone, talk. It's built in, it's free, and for a one-line question there's nothing to gain from a second tool — you'd be adding a step, not removing one. The same goes for the voice features inside the other big chatbots when you're working only in that one app.

A dedicated hotkey pulls ahead in three spots: long prompts you want to read before sending, the filler cleanup, and using the same dictation habit across every app instead of just Gemini. The more your prompts look like real writing — context, instructions, a question — and the more tools you bounce between, the more those matter.

Pick the smallest tool that solves your problem. For one spoken question, that's the mic already in Gemini. For long prompts dictated across Gemini and ChatGPT and a half-finished email, the system-wide hotkey stops feeling like overkill around the second paragraph you didn't have to type.

What Whisper costs

The local dictation pipeline — transcription and the AI cleanup over Ollama — is free for any signed-in user, with no card at signup. So dictating prompts into Gemini with Whisper costs nothing. Whisper Pro adds the cloud features (OpenAI transcription, cloud AI enhancement, voice web search), and it carries a separate trial. The exact numbers live on the pricing page rather than here, because prices move and a blog post is a bad place to keep them current.

Gemini can hear you already, and for a quick question that's the end of the story — use the mic that's there. The reason to dictate the prompt into the box instead is the part the mic can't do: you get to read it before it runs, and you get to do it the same way in every other app you talk to. Two tools, two jobs. Most of the time, the smaller one is enough.