← Back to Blog

Short answer: Click into the OmniFocus field you want to fill, then dictate the task name or note body at the cursor. Speak the prose, keep tags, dates and estimates on the keyboard. The note field is where dictation pays off most, because it is the field almost nobody fills in.

Open your OmniFocus database and read ten tasks you wrote more than a month ago. Some of them will be perfectly clear. Some will say things like "Follow up Anderson," "Fix the deck," or "Reply re: the thing from Tuesday." You wrote those with full context in your head and none of it on the screen, and now the context is gone. The task is still on the list because you cannot bring yourself to delete it and cannot remember what it meant.

That is not a discipline failure. It is an input cost. Writing a task that will still make sense in six weeks takes about three sentences, and three sentences take roughly forty seconds to type at the pace most adults actually type. Writing a task that makes sense right now takes four words. When you are mid-meeting, mid-call, or mid-thought, the four-word version wins every time, and the cost lands on your future self during a weekly review.

Dictation changes the arithmetic. Most people speak at 130 to 150 words per minute against roughly 40 words per minute of sustained typing. The three-sentence version stops being expensive. This guide covers how to actually dictate into OmniFocus on both Mac and iPhone, which fields to speak and which to type, and the specific failure modes that make people give up in the first ten minutes.

Why OmniFocus in particular rewards dictation

OmniFocus is built around a structure that most task managers only gesture at. Every item has a name, and every item also has a note field. There are projects, tags, defer dates, due dates, review intervals, and perspectives that slice the database in ways you define. The structure is excellent. The problem is that the structure only works if there is something written in it.

People who have used OmniFocus for years usually end up in one of two states. Either the database is thin, with lots of terse task names and almost no notes, or it is thorough and the person spends a meaningful amount of time every week typing into it. The thin database fails during review because you cannot judge a task you cannot decode. The thorough database works, but it costs an hour a week that nobody enjoys spending.

Dictation gives you the thorough database at the cost of the thin one. The note field in particular is where this compounds. A note that says why the task exists, what "done" looks like, and what you already tried turns a weekly review from an act of archaeology into an act of decision-making.

The core rule: speak the sentences, type the structure

This is the single most important habit, and getting it wrong is why most people conclude that dictation "does not work" with a structured task manager.

Dictate: task names written as full instructions, note bodies, meeting outcomes, the reasoning behind a decision, draft replies, checklists written as prose, anything that is a sentence.

Type: tags, project assignment, defer and due dates, time estimates, repeat rules, flags, and anything numeric.

The reason is not that dictation cannot handle short strings. It is about how errors reveal themselves. If dictation garbles a sentence, you can see that it is garbled. Prose defends itself. If dictation turns a due date of the 13th into the 30th, or drops a digit from an invoice number in a task name, the result reads as perfectly correct and stays wrong until it causes a problem. Structured fields have no redundancy, so an error in them is invisible.

There is a second, more practical reason on the structural side. OmniFocus has its own entry conventions for tags, dates and projects, with autocomplete that fires as you type. Those pickers are designed for keystrokes. Speaking the word "tomorrow" into a date field does not always mean what you expect, and speaking a tag name does not summon the tag picker. Use the keyboard for those, where the interface is already fast, and use your voice for the part where the keyboard is slow.

Setting up dictation on the Mac

OmniFocus on macOS uses standard Mac text fields throughout. That means any system-level dictation writes into it without a plugin, an extension, or anything installed inside OmniFocus itself. This matters more than it sounds. If you sync a database across machines, nothing you do here travels with it, and nothing breaks when OmniFocus updates.

Voice Keyboard Pro sits in the Mac menu bar and works the same way in every app. You hold a hotkey, speak, release, and the text appears at your cursor. There is no window to switch to, no transcript to copy out, and no per-app configuration. If it works in Mail, it works in OmniFocus.

The three-step flow

  1. Put the cursor where the text goes. Click into the task name field, or open the inspector and click into the note field. This sounds obvious and it is the number one cause of "my dictation went nowhere." Text appears at the cursor, so there has to be a cursor.
  2. Hold the hotkey and speak. One thought per pass. Do not try to dictate a name, a note and three tags in one continuous stream.
  3. Release and look. Read what landed before you move on. This takes two seconds and saves you a confusing review three weeks later.

Watch your hotkey for collisions

OmniFocus users tend to accumulate keyboard shortcuts, including a global quick-entry shortcut that works from any app. If you pick a dictation hotkey that collides with quick entry, or with a shortcut that another Mac app has claimed, you will get one of two confusing results: nothing happens, or the wrong thing happens while you are talking.

The reliable fix is to use a key that almost nothing claims. A function key in the F13 to F19 range is ideal on a full keyboard. On a laptop, a single right-side modifier held on its own is a good option. Whatever you pick, test it once inside OmniFocus and once in a plain text editor before you trust it.

The five OmniFocus surfaces worth dictating

1. The note field (highest value, least used)

This is the one that changes how the system feels. A task name is a label. The note is where the task becomes actionable.

A useful note answers three questions: why this exists, what finished looks like, and what has already happened. Spoken, that is about fifteen seconds:

"This came out of the Thursday call. Anderson wants a revised scope before their board meeting, so the deadline is theirs, not ours. Done means a two-page scope doc sent to Anderson with the new phase two pricing. I already have the old scope doc; what is missing is the pricing table."

Typed, that is roughly forty-five seconds of work you will skip. Spoken, it is one held hotkey. Six weeks later, that note is the difference between doing the task and deferring it again.

2. Inbox capture that is actually a full thought

Quick entry exists so that capture is cheap. The problem is that cheap capture produces items you have to process twice: once to understand what you meant, once to decide what to do. Dictating into quick entry lets you capture a full sentence at the same speed you would have captured a fragment.

Instead of "email Priya," capture "Email Priya to ask whether the vendor contract needs legal review before we sign, because the renewal auto-triggers at the end of the month." That item can be processed once and filed. If you want a broader look at the capture habit itself, our post on dictation for brainstorming and capture covers the same idea outside a task manager.

3. Project notes and the reason a project exists

Most people write a project name and nothing else. Six months in, they have a project called "Website" with nineteen tasks and no stated outcome, which makes it impossible to decide whether it is finished.

Dictate a short project note when you create the project: what outcome closes it, who cares about it, and what the constraint is. This is thirty seconds of speech that saves an entire agonised review session later. It also makes the difference between a project that is genuinely stalled and one that only looks stalled because you forgot the point.

4. Meeting outcomes, right after the meeting

The five minutes after a meeting ends is when you know the most and remember it best, and it is also when you have the least appetite for typing. Dictating the outcomes straight into the relevant project turns a vague "I should write that up" into three concrete tasks with real notes attached.

On a Mac, Voice Keyboard Pro also has Meeting Mode, which does speaker detection and AI notes for a call, and calendar meeting detection so it can pick up that a meeting is starting. That is useful for the record of what was said. It does not replace the step described here, which is the judgement call about what you are going to do next. The transcript is raw material; the OmniFocus entry is the decision. If you record or take AI notes on a call with other people, tell them.

5. The weekly review, out loud

Review is the part of the OmniFocus method that people abandon first, and the reason is usually that it is passive. You look at a list and feel vaguely bad. Making review an active writing exercise changes it.

As you go through each project, dictate one sentence into the project note: what moved, what is blocked, what the next real action is. It adds maybe ten minutes across a whole review and it means next week's review starts with information instead of a blank stare. This is the same principle as an end-of-session summary in research work, which we cover in the post on dictating in Zotero.

Dictating into OmniFocus on iPhone

Mobile is where most capture actually happens, and it is where typing is slowest. It is also where the terse-fragment problem is worst, because typing a paragraph on a phone while walking is genuinely unpleasant.

Voice Keyboard Pro on iPhone is a custom keyboard with a built-in mic button, so it works in any app that accepts a keyboard, including OmniFocus. You switch to it the way you switch to any keyboard, tap the mic, and speak. The text goes into whatever field you had focused.

Three things make mobile capture work well:

One note on third-party keyboards generally: some apps and some managed work devices restrict them. If the keyboard does not appear in OmniFocus, check the keyboard settings and Full Access before assuming something is broken. Our guide on why a third-party keyboard keeps switching back walks through the usual causes.

Smart Vocabulary and the proper-noun problem

Here is the thing that decides whether you keep using dictation with a task manager after week one.

Your OmniFocus database is full of proper nouns. Client names, colleague names, product codenames, internal system names, the odd acronym your company invented. General-purpose transcription is good at English and has no idea that your project is called Halcyon or that your colleague spells it Sian rather than Shawn. Every time it gets one wrong, you stop, reach for the trackpad, and fix it, and after the fifth interruption you conclude dictation is more trouble than it is worth.

Smart Vocabulary is the fix. It is a personal dictionary with replacement rules, so you tell Voice Keyboard Pro once how your recurring names are spelled and it stops getting them wrong. Twenty entries covers most people's working vocabulary: your top ten clients or colleagues, your active project names, and any internal acronym you say weekly.

This matters more in OmniFocus than in a document, for a subtle reason. Search and perspectives depend on consistent spelling. A task about "Halcyon" and a task about "Halcion" do not show up in the same filtered view. Inconsistent proper nouns quietly fracture your database. Our post on custom vocabulary that learns your words goes deeper on how to build the list.

Common problems and what causes them

Nothing appears when I dictate

Ninety percent of the time there is no cursor in a text field. Click directly into the task name or note body first. The rest of the time it is a hotkey collision, or another app is holding the microphone. On a Mac, the orange dot in the menu bar tells you when something else has the mic.

My text lands in the wrong field

OmniFocus moves focus as you work, and the inspector is a separate surface from the outline. If you dictate right after pressing Return or Tab, focus may have moved. The habit that fixes this permanently is: click, then look at where the caret is blinking, then speak.

Dictation adds a task instead of typing into one

If Return is bound to "create new item" in the context you are in, any spoken newline or trailing punctuation behaviour can commit the row early. Dictate the text first, then press Return yourself when you actually want the next item. Never rely on speech to trigger a structural action.

My dates come out wrong

This is the reason for the speak-prose, type-structure rule. Do not dictate into date fields. Type them. A wrong date does not look wrong.

The transcription is noticeably worse in some rooms

Background noise and mic distance matter more than most people expect. Speaking closer to the mic at a normal volume beats speaking loudly from across a desk. Our guide on fixing dictation in noisy environments covers the practical setup.

How this compares to other task managers

The principle is the same everywhere: speak the content, type the structure. What differs is how much structure the app has and how much content it wants.

Lightweight apps like Apple Reminders have less structure to type around, so a bigger share of what you enter can be spoken. We cover that in the posts on dictating in Apple Reminders and dictating in Things. OmniFocus sits at the heavy end. It has the most structure, which means more keyboard work in the fields, and it also has the richest note field, which means dictation has the most room to add value. The ratio is different, the rule is not.

A realistic first week

Do not try to convert your whole workflow at once. Here is a sequence that tends to stick.

Honest limits

A few things worth saying plainly.

Dictation does not organise anything. It puts text where your cursor is. Projects, tags, review intervals and perspectives are still yours to design, and a badly designed OmniFocus database with beautifully dictated notes is still a badly designed database.

It needs a connection. Voice Keyboard Pro's transcription engine runs in the cloud, so dictation is not available on a plane without wifi. Keep the built-in system dictation as a fallback if you work offline often.

Your voice is audible. In a shared office, the notes you dictate can be overheard, which is a consideration if your task database contains client-confidential material. We wrote about the practical version of this in the post on dictating in an open-plan office. The risk there is the person two metres away, not the network.

On privacy: Voice Keyboard Pro's server stores only operational pings. No audio and no transcript content is retained. If your task database holds material under a confidentiality obligation, read our policy yourself rather than taking a sentence in a blog post as a control.

The point

OmniFocus is a system for making decisions about your commitments. It only works if the commitments are written down in enough detail to decide about. For most people, the gap between the database they have and the database that would actually work is not a knowledge gap or a discipline gap. It is a few hundred sentences that were too expensive to type.

At speaking speed, those sentences cost about fifteen seconds each. Voice Keyboard Pro has a free tier, and Pro is $4.99 a month or $34.99 a year. Try it on the next ten tasks you capture. Fill in the note field for each one. Then run your weekly review and see how different it feels to review a list you can actually read.