Voice to text in Obsidian

Does Obsidian have built-in voice-to-text?

No. Obsidian has no native dictation feature that turns your speech into text in the editor. The core Audio Recorder plugin records an audio file and attaches it to your note; it does not transcribe what you said into words. That's a useful thing, but it's the opposite of what most people searching for this want.

To get real transcription inside a vault, you reach for a community plugin. The most common is the "Whisper" plugin by nikdanilov. It records or uploads audio, sends it to a Whisper-compatible transcription API, and pastes the result at your cursor. It works on desktop and mobile, and the hotkey is Alt+Q by default. The catch is in its own quick-start: you add an API key in the settings, because it talks to OpenAI, Groq, Azure, or another Whisper-compatible service. That means cloud, and it means paying the API provider per minute of audio.

There are other in-editor routes too. People pair the core Audio Recorder with a transcription service like Deepgram, or run separate dictation tools alongside Obsidian. All of them are fine choices. But notice the common thread: the ones that transcribe inside the note almost all route your audio to a cloud API and ask for a key. There's a different way to do this that skips both.

A plugin in one vault vs. a hotkey everywhere

This is the distinction worth getting straight. An in-vault plugin lives inside Obsidian. It transcribes when you're in a note, using its own hotkey, and it only works there. A system-wide hotkey app lives one level up, in your operating system. It types into Obsidian, but also into your browser, your email, your terminal, and every other app, because as far as your computer is concerned, you're just typing.

Which one you want depends on what you're actually doing. If your only goal is to drop a recorded audio file into a note and have it transcribed in place, the plugin is built for exactly that. If you want to dictate prose into the editor the way you'd type it, and also dictate everywhere else you write, the system-wide app is the better shape. The rest of this guide is mostly about the second kind, because it's the one that needs no key and no per-vault setup.

The fastest way: a system-wide hotkey

Here's where a desktop app changes the math. Whisper by Remskill is not an Obsidian plugin, a vault integration, or a browser extension. It's a desktop app that works like a keyboard: press a hotkey, speak, and the transcript is pasted at the cursor, in any app, your Obsidian note included. There's no API key to paste, because the transcription engines run locally on your machine. It works the same in Obsidian and in everything else, since it types rather than integrates.

Setup is short:

That's the whole loop. The transcript appears in the editor, you read it, you keep writing.

Speak, and the note fills in

Once it's running, the experience is unremarkable in the best way. You put your cursor in a note, hold the key, talk, let go. A second or so later the text is sitting in the editor as if you'd typed it. No separate recording to scrub through later, no audio file cluttering the vault, no transcription job to wait on. Just words where your cursor was.

Because the local transcription runs on your machine (pure-Rust, no Python sidecar, no server in the loop), it works offline. For a notes app, that's the whole point. The thought you're capturing, the meeting you're summarizing, the half-formed idea you don't want indexed by someone else's logs, none of the audio leaves your laptop. The in-vault plugins that call a cloud API can't say that, and Windows' own built-in voice typing needs an internet connection to work at all. Whisper's local mode does not.

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 note. It's there when you want to tune.

Cleaning up dictated notes with AI

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, yeah, so the, the meeting moved, I think we're, we're doing Thursday now" becomes "The meeting moved; I think we're doing Thursday now." For a permanent note you'll re-read later, that cleanup earns its keep.

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 require Pro; it's there in the free local pipeline. You can also turn it off and paste the raw transcript, which for a quick capture is often exactly right.

Local and private: your notes stay on your machine

This is the one strong opinion I'll put my name on for a notes tool: cloud-only dictation is a privacy disaster waiting to be transcribed. Your vault is where you keep the salary spreadsheet you're drafting, the email to your kid's school, the half-thought you're not ready to share. The whole appeal of Obsidian for a lot of people is that the notes are plain files on their own disk. Piping the audio for those notes through a vendor's API undoes that quietly.

With Whisper's local engines, the recording is transcribed on your own CPU and discarded. No key, no account at the transcription provider, no audio sitting in a log somewhere. That's not a marketing line, it's just where the computation happens. I once watched a team rack up a five-figure cloud bill in a quarter feeding standup recordings to an API; the CFO's response was, roughly, "or we don't pay to transcribe meetings that already have notes." The same logic applies to a personal vault, minus the CFO.

When a vault plugin is the better fit

I'm not going to pretend the system-wide app wins every case, because it doesn't. If what you actually have is a folder of recorded voice memos or meeting audio that you want transcribed and attached inside specific notes, an in-vault plugin like the nikdanilov "Whisper" plugin is built for exactly that workflow. It records or uploads a file and drops the transcript into the note. You'll pay the transcription API per minute and manage a key, but the file-in, transcript-in-note loop is what it does, and a hotkey-typing app doesn't replace it.

There's also the do-nothing option. If you only dictate the occasional line, your operating system already does it for free. On Windows, press the Windows logo key + H and voice typing opens in any text box, an Obsidian note included. On a Mac, the built-in Dictation shortcut does the same, and on Apple silicon it runs on-device with no internet. The tradeoff: Windows' Win+H needs a connection to run, while macOS dictation and Whisper's local mode don't.

Pick the smallest tool that solves your problem. For a one-line capture, that's the key your OS already has. For transcribing audio files into specific notes, that's an in-vault plugin. For dictating prose into the editor and everywhere else you write, offline and without a key, that's the system-wide app.

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 getting your voice into a note with Whisper costs nothing, and there's no transcription API metering you per minute. 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.

Obsidian will probably grow a native dictation feature eventually, the way apps tend to once enough people install three plugins trying to fake one. Until then, your vault has options: an in-vault plugin for transcribing audio files, your OS for the occasional line, and a system-wide app for dictating prose into the editor without a key or a server in the loop. Three tools, one notebook. Most of the time you need the one that keeps the audio on your machine.