How to Set Up Voice Dictation on Your Mac in 10 Minutes

Getting voice dictation working well on a Mac is a 10-minute job if you know what to change. Most guides skip the settings that matter and leave you with mediocre accuracy. This one does not.
What You Need Before You Start
Your Mac's built-in microphone will work, but a dedicated microphone makes a noticeable difference in accuracy. You do not need to spend much. A $30 to $50 USB cardioid microphone like the Blue Snowball or the Samson Go Mic outperforms the built-in mic in most real-world environments.
If you are at a desk, a small desktop microphone positioned 6 to 12 inches from your mouth is the sweet spot. Too close and you get plosives on P and B sounds. Too far and background noise competes with your voice.
AirPods and other Bluetooth headsets work but introduce a small latency that some people find distracting. A wired option is more consistent.
Installing VoiceInk
Download VoiceInk from the Mac App Store or directly from the VoiceInk website. It installs like any other Mac app. On first launch it will ask for microphone permission and accessibility permission. Both are required. The accessibility permission is what allows it to type into any app, which is the core functionality.
All processing happens on your device. Nothing you say is sent to a server. That matters if you are dictating anything sensitive.
Setting Your Trigger Key
The default trigger is a hold-to-talk key. You hold it, speak, release, and the text appears. You can change this key in VoiceInk preferences to whatever fits your setup.
Common choices are the right Option key, the Caps Lock key, or a mouse thumb button if your mouse has one. The ideal choice is something you can hold without tension, a key your thumb or non-dominant hand can reach comfortably without your hand moving far from its resting position.
Avoid repurposing a key you use frequently for typing. The goal is a dedicated, low-conflict shortcut.
Improving Accuracy From Day One
A few things have a large effect on accuracy:
Speak at a consistent pace. You do not need to speak slowly, but erratic rhythm, normal speed then very fast then trailing off, confuses the model. Find a pace and stay there.
Speak clearly at the start and end of sentences. These are the points where transcription errors cluster. Do not drop your voice at the end of a sentence.
Reduce background noise where you can. An air conditioning vent pointing at your microphone, a fan, keyboard noise, these all degrade accuracy. A small acoustic change in your environment has a larger effect than most software settings.
If you use proper nouns, technical terms, or domain-specific vocabulary frequently, dictate a test paragraph with those terms a few times in the first session. The model adapts to your voice and common vocabulary quickly.
Testing Your Setup
Before you use VoiceInk for real work, do a two-minute test. Open a blank document in any app, trigger VoiceInk, and read two or three paragraphs from something you have nearby, a book, an article, anything.
Look at the output. If accuracy is above 95 percent, your setup is good. If you are seeing consistent errors on specific sounds or words, check your microphone distance first, then your room noise, then whether your input source is set correctly in System Settings under Sound.
Day-One Workflow Suggestion
Do not try to replace all typing immediately. Start with one category of task: just emails, or just meeting notes, or just first-draft writing. Do that for a week until it feels natural.
Speed comes with practice. The first session will feel slow. The fifth session will feel normal. The twentieth session will feel like the only sensible way to produce text.
Your setup is done in 10 minutes. Everything after that is just using it.
Stop typing. Start talking.
VoiceInk turns your voice into text in any app. Local, fast, private. Free to start.
Download VoiceInk Free