Short answer: Course creators should dictate lesson scripts, module descriptions, worksheets and student replies, and keep pricing, timestamps and quiz answer keys on the keyboard. A spoken script also sounds like teaching rather than an essay read aloud, which is the real reason to dictate it.
Most unfinished online courses are not unfinished because the video is hard. The video is the part people look forward to. They stall on everything wrapped around the video: the script, the lesson description, the workbook, the quiz questions, the welcome email, the sales page, and the growing pile of student questions in the inbox.
All of that is text. And text, for most people, arrives at roughly 40 words per minute, which is where the average adult types. Speech runs at 130 to 150. Even a fast professional typist working at 80 to 100 words a minute is producing text at well under speaking speed, and a course is not a document. It is dozens of documents that all have to exist before anyone can buy the thing.
This guide is about closing that gap using voice, with one non-obvious benefit that matters more than the speed: a script you speak reads like teaching, and a script you type reads like an essay being read aloud. Students can hear the difference immediately.
Count the writing in one course
It is worth listing this out, because creators consistently underestimate it. A single modest course typically requires:
- An outline and a module map
- A description for every module and every lesson
- A script or a detailed talking-point sheet per video
- Slide text and on-screen callouts
- Worksheets, templates and handouts
- Quiz or assessment questions, plus feedback text for wrong answers
- A sales page
- A welcome email and an onboarding sequence
- Launch and re-engagement emails
- Community or comment posts, week after week
- Replies to individual student questions, forever
- Captions, transcripts and alt text for accessibility
The video is maybe a fifth of the work. Everything else is a typing job, and it is the reason the second module has been "almost done" since March.
The real argument: spoken scripts sound spoken
Start here, because it is the part that changes the product rather than just the schedule.
When you write a lesson script by typing, you write in your writing voice. Written prose is denser, more formal, and more complete than speech. Every clause is finished. Nothing is repeated. There are no asides. Read that aloud to a camera and it sounds like a person reading, because it is.
When you speak a lesson script, you produce teaching language. You repeat the important thing twice in different words. You put the example before the definition because that is how you would actually explain it to someone. You say "and here is where people usually get stuck," which is a sentence almost nobody types into a script and almost everybody says out loud.
So the method is not "dictate your script to save time." It is: teach the lesson out loud to an imaginary student, capture what you said, then edit it into a script. The time saving is real, but the tonal improvement is the reason to do it.
How to actually do it
- Open an empty document and put one line at the top: what the student should be able to do after this lesson.
- Stand up. Talk through the lesson as if one specific person is in front of you, dictating as you go. Do not stop to fix anything.
- Stop when you have run out of things to say. The result will be too long and repetitive.
- Now edit with the keyboard. Cut the false starts, keep the good asides, and put the whole thing in order.
The second pass is fast because you are deleting instead of inventing, which is a much lighter cognitive task. This is the same two-pass structure that works for first drafts of long-form writing: produce something bad quickly, then apply judgment to something that exists.
Module and lesson descriptions
Descriptions are where courses look thin. Nearly every creator writes a strong description for module one and increasingly terse ones after that, and the reason is purely mechanical: by module six you have typed five of them and you are tired of the form.
Dictate all of them in one sitting instead. Say, for each lesson: what it covers, who it is for, and what changes for the student afterwards. Three sentences, fifteen seconds each. Doing them consecutively also produces a side benefit that is hard to get any other way, which is that you notice the overlaps and gaps in your own curriculum while you are describing it. Two lessons that turn out to have the same description are two lessons that should probably be one.
Worksheets, prompts and assessments
Worksheets are the most-skipped deliverable in course building and the one students most often cite as the reason a course felt worth it. They are also almost entirely prose: an instruction, a prompt, a couple of guiding questions, a space to write.
Speak the prompts. They come out sounding like a person asking a question rather than a form.
The one boundary to hold here is assessments. Speak the question stems and the explanation text. Type the answer keys, the option letters, the point values, and anything numeric. This is the general rule that applies across every kind of dictation work: prose defends itself and data does not. A clumsy sentence looks clumsy, so you fix it. A quiz answer key where B became D looks exactly as correct as the right one, and the first person to notice is a student who got a right answer marked wrong.
The same split covers your pricing, your refund window, your cohort dates and your timestamps. Say the sentences, type the numbers, verify the numbers.
Student replies, which is where the evenings go
Every established course creator arrives at the same place: the course is built and the inbox is now the job. Questions in the community, questions by email, comments under lessons, people asking the thing that module three already answers.
Two things make this manageable, and both depend on writing being cheap.
The first is answering properly the first time. A thorough answer takes four or five sentences. Typed, that is a couple of minutes and you will not do it forty times, so you write "see module 3" and the student comes back unsatisfied. Spoken, it is twenty seconds, and the full answer is the one that actually closes the thread.
The second is reuse. Keep a document of your best answers. Once you notice a question repeating, dictate a clean canonical version into that file and paste it from then on, with a sentence or two of personalization spoken fresh each time. Over a year of a live course this compounds into your best FAQ, your best sales page objection handling, and often your next lesson, because a question three people asked is a gap in the curriculum.
Voice Keyboard Pro works inside whatever app the conversation is happening in. On the Mac you hold a hotkey, speak, and release, and the text appears at your cursor in the browser field, the email client or the community platform, with nothing to install into those tools. On the iPhone it is a keyboard with a mic button built in, which is what makes replying to a student while away from the desk realistic rather than theoretical.
Sales pages and email sequences
Launch copy has the same problem as lesson scripts, in a sharper form. Typed sales copy tends toward the generic, because the blank page invites you to write what a sales page is supposed to sound like. Spoken copy tends toward the specific, because when you say out loud who this is for, you name an actual person you have taught.
A workable approach: dictate answers to five questions before you write any page copy at all. Who is this for. What are they stuck on. What will they be able to do afterwards. Why should they believe you. What happens if they do nothing. Speak two or three minutes on each. The raw output is not a sales page, but every good line on the eventual page will come out of it, and you will have produced it in fifteen minutes rather than over a weekend.
The same applies to your email list. If you also run a newsletter alongside the course, the habits in our guide to dictating email newsletters transfer directly.
Smart Vocabulary for your course's own language
Every course develops proprietary vocabulary: your framework name, your named steps, your acronym for the method, the product names of the tools you teach, and the names of students you reply to often.
These are exactly the words general transcription is worst at, because they are invented or used unusually. Correcting the same framework name thirty times a day is the single most common reason people give up on dictation before it becomes a habit.
Voice Keyboard Pro's Smart Vocabulary is a personal dictionary with replacement rules. Add your framework terms, your module names, the tools in your stack and the names you type most, and they land correctly from then on. Build the list from real corrections after a couple of days of use rather than trying to guess it up front.
Office hours and group calls
If your course includes live calls, coaching sessions or office hours, that is a large amount of content you are currently producing and then discarding. Meeting Mode on the Mac handles the recording side with speaker detection and AI notes, and calendar meeting detection means it can recognize a scheduled session rather than relying on you to remember.
Two things to be clear about. Tell people on the call that you are capturing notes, every time, as a matter of basic courtesy and because the norms and rules around recording vary by where you and your students are. And treat the output as raw material rather than a deliverable: the useful artifact is the summary you write afterwards, not a transcript dumped into the community.
On iPhone: capture, edit, and international students
Course ideas do not respect the desk. The clearest explanation of the thing you have been struggling to teach usually arrives while walking, and by the time you are back it has degraded into a topic rather than a paragraph.
The iPhone keyboard puts dictation inside any app, so you can capture the explanation in whatever notes app you already use, in full sentences, in the ten seconds where it still exists. Voice Edit lets you speak a change to text you have already dictated, which is much faster than thumb-editing a paragraph.
Two-way translation while dictating covers 24 languages and is genuinely useful if you have international students. A student writes in Spanish, you reply in English and it goes out translated. Treat translated output as a strong first draft, especially for anything that states a policy, a refund term or a date. If it is money or a deadline, have someone who speaks the language read it before you send it.
A batching workflow that works
Do not spread this across the week. Dictation rewards batching, because the expensive part is loading the context and the cheap part is producing the words.
- One session for descriptions. Every module and lesson description, back to back.
- One session per lesson script. Talk it through once, edit later in a separate sitting so you are not editing and generating at the same time.
- One session for the workbook. All prompts across all modules in a row, which also keeps their tone consistent.
- A standing 20-minute slot for student replies. Answer thoroughly, and file anything that repeats into your canonical answers document.
Honest limits
Voice Keyboard Pro types text at your cursor. It does not write your course, structure your curriculum, record or edit video, or generate captions for your lessons. It does not integrate with any course platform, which is the point: it works in all of them because it works below all of them, in the ordinary text fields your browser and apps already have.
It needs a connection. It runs on macOS and iOS. And it is audible, which is worth planning around if you share a room, because dictating a reply to a student means the room hears the reply.
On privacy, our server stores only operational pings. No audio and no transcript content is stored server-side. If your course handles anything sensitive about your students, read the policy yourself and make your own decision rather than relying on a summary here.
The second module has been almost done since March because writing it costs three times what recording it costs. Change the cost and the backlog moves.
Start with one lesson
Pick the lesson you have been avoiding. Stand up, talk it through to an imaginary student while dictating, and stop when you have nothing left to say. You will have a messy transcript that is closer to a finished script than the blank page you have been staring at for weeks, and you will have made it in the length of the lesson itself.
Voice Keyboard Pro has a free tier with daily limits, and Pro is $4.99 a month or $34.99 a year. If you want a broader look at how independent workers restructure their writing around voice, our guide for freelancers covers the same ground from the client-work side.
FAQ
Does this work with my course platform?
It works in any ordinary text field, which is where course platforms put their lesson editors, description boxes, quiz builders and community replies. There is nothing to install into the platform and no integration to maintain, so switching platforms later changes nothing about your workflow.
Should I dictate the script or just record without one?
Dictate a rough script, then edit. Recording with no script tends to produce lessons that ramble and are painful to edit. Recording from a typed script tends to produce lessons that sound read. Dictating and then trimming gives you structure and spoken tone at the same time.
What about captions and transcripts for my videos?
That is a different job. Voice Keyboard Pro writes text where your cursor is rather than processing an existing video file. For captions, use your platform's own captioning, then fix the terminology by hand.
How do I stop it from mangling my framework name?
Add it to Smart Vocabulary. Invented names, acronyms and unusual spellings are the words transcription struggles with most, and one entry fixes it permanently.
Can I answer student questions from my phone?
Yes. The iPhone keyboard has a mic button built in and works in mail, browsers and community apps, so a thorough reply is a twenty-second job rather than a five-minute thumb-typing job.
Is dictated writing worse than typed writing?
Different, not worse, and for teaching material it is often better. Dictated prose is looser and more conversational, which is a liability in a legal contract and an asset in a lesson. The fix for the looseness is a fast editing pass, not typing the first draft.