TL;DR: Voice typing for Obsidian should match the reason you chose Obsidian: local files you own, no cloud dependency. Infina is that dictation app: transcription runs entirely on your Mac, audio never leaves your device, and it is a flat $99/year with unlimited words instead of a monthly bill for cloud transcription. And it does something no plugin or hotkey app does: with hands-free mode on, you say "open Obsidian", then "type" plus your thought, then "send", and the whole capture happens by voice from across the room, no keyboard touched. 7-day refund if it's not for you.
Why voice typing for Obsidian is a values question
Obsidian users didn't pick their tool casually. You chose plain-text Markdown files, on your own disk, readable in fifty years, hostage to no company. Local-first isn't a feature to you; it's the premise.
So it's strange to bolt on a dictation tool that ships every half-formed thought to someone else's server, and goes dark the moment that server is unreachable.
Your dictation app should hold the same beliefs your notes app does. That means transcription on your machine, audio that never leaves it, and no round trip to anyone's cloud.
That's Infina. By default it transcribes your speech entirely on your Mac using an on-device model on Apple Silicon. It works offline, on a plane, in a dead zone, anywhere your vault works. And it's a flat $99/year with unlimited words, not $15 a month. The full privacy story is in on-device dictation for Mac.
What about the Obsidian Whisper plugin?
Worth addressing honestly, because it's the first thing Obsidian users find.
As of July 4, 2026, community plugins like Whisper add speech to text inside Obsidian. They work, and the community deserves credit for building them. But look at the shape of the solution:
- Per-app. The plugin dictates into Obsidian only. Your email, your browser, your AI tools get nothing.
- Setup. You install the plugin, get an API key, paste it in, and configure endpoints. Most route your audio to a cloud Whisper API, which is exactly the trip a local-first person picked Obsidian to avoid. Fully offline variants exist but are their own separate setup.
- Maintenance. Community plugins live on volunteer time and can break with app updates.
Infina inverts all three. It types at the OS level, so it works in Obsidian and every other app on your Mac with zero plugins and zero API keys. Hold Option (⌥), speak, release, and the text lands at your cursor: on-device, offline-capable, nothing to configure.
One tool for your whole Mac beats one plugin per app. If part of your system lives in Notion or you draft in Google Docs, the exact same gesture works there too.
Daily notes and zettelkasten capture by voice
The workflows Obsidian people actually run are capture-heavy, and capture is where dictation shines.
Daily notes. Open today's note, hold Option, and narrate: what you're working on, what's blocking you, the thing you promised someone. Three sentences in ten seconds. The daily note habit survives on low friction, and voice is the lowest friction there is.
Zettelkasten and fleeting notes. The whole method depends on catching fleeting thoughts before they vanish, then refining them into permanent notes later. Speak the fleeting note the moment it occurs. Refine at the desk, on your schedule.
Literature notes. Book in one hand, speak your reaction into the note. No putting the book down to type.
Base Infina outputs raw text on purpose: fast on-device transcription with rule-based cleanup, no rewriting, no cloud round-trip. For a vault, raw is right. These are your notes, in your files, and messy-but-captured beats polished-but-lost every time. Plain speech into plain text: it's the most Obsidian-shaped pipeline imaginable.
When you do want polish, dictating a post or something others read verbatim, the optional cloud add-on adds sharper cloud transcription plus large language model cleanup for $9.99/month on top of your plan. That's the same polish cloud dictation apps charge $15/month for as the only way to use them at all; with Infina it's optional, cancel anytime, and the app falls back to on-device the moment you do.
The hands-free capture loop
Here's the part no plugin, no built-in dictation, and no cloud dictation app does.
Every other dictation tool assumes your hands are at the keyboard: a hotkey press or a mic click for every single capture. Fine at the desk. Useless when the idea arrives while you're making coffee.
Infina has a hands-free mode. Double-tap Cmd (⌘) to turn it on, and the whole loop runs by voice from 2 to 3 feet away:
- Say "open Obsidian". It comes to the front.
- Speak a sentence starting with "type": "type the zettelkasten failure mode is capture friction, not linking." Infina types it into the focused note.
- Say "send" when you want Enter pressed, like confirming a new line or a quick switcher entry.
- Say "open Notes" or "open Notion" to capture somewhere else, then loop back.
Pacing while thinking through an essay, hands in the sink, book in both hands: the thought still lands in the vault. While hands-free waits for you, listening runs on-device too; nothing is recorded or sent anywhere.
Honest caveats: hands-free ships off by default and is labeled experimental, and it likes a reasonably quiet room. Hold-Option push-to-talk is the always-on fallback.
Local-first on both sides
The economics rhyme with the philosophy.
Obsidian's core app is free for personal use, and your files are plain Markdown on your own disk regardless. Infina is $99/year (price at the time of writing): flat, with unlimited words and unlimited voice commands, every update included, and a 7-day money-back guarantee, no questions asked.
Compare a $15/month cloud dictation subscription: $180 a year, with every sentence you speak routed through someone else's servers. Infina is $99 a year and transcribes on your Mac. The optional cloud add-on is $9.99/month with its own 7-day free trial, only if and when you want it. Full details on pricing.
A vault that lives on your disk deserves a dictation app that runs on your Mac.
FAQ
Does Obsidian have built-in voice typing? No. Obsidian has no native dictation as of July 4, 2026. Options are community plugins (per-app, setup required, usually cloud API keys) or a system-level dictation app like Infina that types into Obsidian and every other Mac app with nothing to install inside Obsidian.
Is the Obsidian Whisper plugin good enough for dictation? It works for Obsidian-only dictation, but most setups send your audio to a cloud Whisper API with your own key, and it does nothing outside Obsidian. Infina transcribes on-device by default, needs no API key, and works system-wide with one gesture.
Does my audio leave my Mac when I dictate into Obsidian with Infina? No, not by default. Transcription runs entirely on your Mac on the Apple Neural Engine, and it works fully offline. Cloud processing exists only as an optional $9.99/month add-on that you can ignore or cancel anytime.
Can I dictate my Obsidian daily note without touching the keyboard? Yes. With hands-free mode on, say "open Obsidian", then speak a sentence starting with "type" and your words are typed into the focused note; say "send" to press Enter. The whole loop runs by voice from across the room.
Is dictated text clean enough for a zettelkasten? For fleeting and daily notes, yes; raw fast capture is the point, and you refine into permanent notes later anyway. If you want publish-ready prose, the optional cloud add-on applies large language model cleanup for $9.99/month, the same polish cloud dictation apps charge $15/month for.
What does Infina cost compared to dictation subscriptions? $99/year at the time of writing, flat, with unlimited words and voice commands, a 7-day money-back guarantee, and every update included. Cloud dictation apps run around $15/month, which is $180 a year, and they transcribe on their servers instead of your Mac.
The bottom line
You chose Obsidian because your notes should be yours: local, plain, permanent. Your dictation should meet the same bar, and most of it doesn't; it's per-app plugins with API keys, or cloud subscriptions that route your own voice through someone else's servers.
Infina matches the vault's values: on-device transcription, audio that never leaves your Mac, offline-capable, a flat $99/year with unlimited words. Raw speech into plain text, in Obsidian and every other app, with the hands-free "open Obsidian, type, send" loop nothing else offers.
Try it risk-free for 7 days. Your fleeting notes will stop fleeing.