Home / Dictation
Talk. It types.
Nothing leaves your Mac.
Press ⌥⇧D in any app — or hold it like a walkie-talkie — and a live transcript streams into a popup at your cursor. Apple's on-device speech engine does the listening, so no audio and no text ever go to a server. Then teach it the names it keeps mangling, and send the transcript wherever it needs to go.
“Let's move the Mackenzie review to Thursday and loop Priya in on the pricing deck.”
the transcript streams in as you speak — then you pick where it lands
One shortcut, anywhere you can type.
There's no window to open and no app to switch to. Dictation happens where your cursor already is.
Press, or hold
⌥⇧D by default. Prefer a space bar combo? ⇧Space, ⌃Space, and ⌘Space are offered too — pick whichever one your hands already know.
Watch it land
The transcript streams into a popup at your cursor while you're still talking. It stays up when you click into another window, so you can read your notes mid-sentence.
Send it somewhere
Copy it, keep it as a clip, send it to Notes, or write it to a file. Or set up a pipeline once and skip the chooser entirely.
Hold it for a sentence. Toggle it for a page.
Long dictation and quick bursts want opposite things from a shortcut, so Pickclip does both — pick one in Settings, and it applies on the very next press.
Press to start, press to stop
For the paragraph you've been putting off. Your hands come straight back to the keyboard while you talk, and nothing is riding on you keeping a key down.
Hold to talk, release to stop
For the short burst that is most of daily use — a Slack reply, a search query, a prompt — where entering and leaving a mode costs more than the speech does.
Brush the key by accident and it doesn't hand you an empty two-second transcript: a tap under a third of a second keeps recording instead, and the next press stops it.
It learns the words a cloud service would most like to have.
Every dictation tool mangles the same things: your colleagues' names, your product names, your team's jargon and ticket ids. Pickclip lets you fix them — in two different ways, because they fail differently.
Vocabulary words
Hand the recognizer the terms you actually use and it tilts its own decoding toward them — so the fix arrives with the right spelling, inflection and spacing. Reach for this when the engine sometimes gets a term right.
Mackenzie · Slideform · PROJ-1420Replacement rules
A deterministic heard → written fix applied to the finished transcript. Reach for this when the engine always produces the same wrong thing. Whole words only, so a rule for “cat” never touches “category” — and possessives still work.
mackenzie → McKenzieStart from your own clipboard, not from an empty list
The proper nouns in your clipboard history are precisely the terms dictation gets wrong — names, product names, ticket ids, domain jargon, copied all day long. Pickclip can mine recent clips and offer them as suggestions you confirm. Nothing is ever added silently.
The vocabulary that would be most useful to a cloud dictation service is exactly the vocabulary you'd least like to upload.
A transcript isn't the end of the job.
Most dictation apps have exactly one answer for what happens next: they type it into the field you were in. Pickclip catches the transcript first — and then it's ordinary Pickclip material.
Pick a destination
Stop, and the popup offers four: Copy, Save as Clip, Send to Notes, Save as File… — and what the popup shows you is exactly what gets saved.
Or never see the chooser
Configure the steps once in Settings → Dictation and every finished transcript runs through them in order — copy it, paste it back into the app you were in, or run one of your actions on it first.
It's a clip like any other
Save the transcript to your history and everything else in Pickclip works on it — the same actions you run on anything you copy, sending their results wherever you told them to.
Speak a rough thought, get a clean sentence in the reply box. If a step fails, the rest are skipped and the chooser opens with your transcript — nothing is lost.
Your voice is none of the cloud's business.
Pickclip dictates with the on-device speech engine built into macOS 26 — the same silicon that runs the rest of the app's AI. macOS downloads the speech model from Apple once, and from then on your audio is transcribed on your own Mac. No audio and no text leaves the machine. There's no account, no API key, and no per-minute meter counting your sentences.
Ten dollars, once, instead of a bill every month.
The good cloud dictation apps are genuinely good. They're also a subscription, an account, and a copy of everything you say sitting on someone else's server.
| Subscription dictation, the usual shape | Pickclip | |
|---|---|---|
| Price | A monthly bill, commonly $12–15, for as long as you use it | $10 once — the whole app, dictation included |
| Where your voice goes | Uploaded to a server to be transcribed | Nowhere. It's transcribed on your Mac |
| Free tier | Usually capped — so many words a week, then it stops | 14 days, uncapped, then a license. No word limits either way |
| Account | Required, with a sign-in on every machine | None. A license key that works offline |
| Your private vocabulary | Uploaded along with everything else | Stays on the Mac — it only ever reaches the local recognizer |
| What else it does | Dictation | Also a clipboard manager with searchable history, on-device AI actions, and screenshot OCR |
The things people ask first.
Does it work offline?
Yes, once the speech model is installed. macOS fetches it from Apple the first time you dictate; after that transcription runs on your Mac, network or no network.
Is my audio uploaded anywhere?
No. Audio is transcribed by Apple's on-device speech engine, and neither the audio nor the transcript leaves your Mac. Your personal dictionary stays local too — it's handed to the local recognizer, never to a server.
Which language does it dictate in?
Your Mac's language, when Apple's on-device engine supports it. If it doesn't, Pickclip picks the closest variant of your language, and falls back to English.
Can it transcribe a meeting or tell speakers apart?
No. Pickclip dictates one voice — yours, at your cursor, in the app you're working in. It doesn't label speakers or record a room.
Does it keep a history of my transcripts?
There's no transcript library to browse or search. Each transcript goes where you send it: the clipboard, your clipboard history as a clip, Notes, or a file.
How is this different from macOS's own dictation?
It's a different job. System dictation types into whichever field has focus. Pickclip catches the transcript first, so your dictionary can correct it, an action can rewrite it, and you choose where it lands.
What do I need?
macOS 26 or later on Apple Silicon, and permission to use the microphone. The speech model downloads once from Apple, the first time you dictate.
What does it cost?
$10 once, for all of Pickclip. Free for 14 days with no account; dictation is included in the trial like everything else.
Say the next paragraph instead of typing it.
Dictation comes with Pickclip — along with a searchable clipboard history, on-device AI actions on anything you copy, and screenshot OCR. One download, one price, nothing uploaded.
↓ Download for Mac Free 14-day trial · $10 once · macOS 26+ · Apple Silicon · microphone permission required