When you press a hotkey to start dictating, something interesting happens in the time it takes to blink. Your words transform into text faster than you can finish speaking them. But not all voice-to-text tools work this way. Most feel sluggish. Some require internet. And nearly all add noticeable delay.
The difference comes down to delay — the gap between the moment you speak and the instant text appears. At 80 milliseconds, Dictato on macOS delivers dictation that feels like an extension of your thoughts, not a separate task.
Let’s look at what that number actually means, why it matters for your workflow, and how it’s possible on a Mac.
What does 80 milliseconds actually feel like?
Eighty milliseconds is one-twelfth of a second. That’s faster than a blink.
Research shows that people notice delays above 100 milliseconds — anything below that feels instant. At 80ms, the text appears before your brain has fully registered you’ve started speaking. It’s faster than your reaction time (most people react in 200-250ms).
In practical terms: you finish speaking a sentence, release the hotkey, and the text is already waiting for you. No pause. No waiting. It just feels like the app is reading your mind.
This is the sweet spot for real-time dictation.
Why speed matters for voice-to-text workflows
It might seem like a fraction of a second doesn’t matter. You’re already speaking, which is slower than thinking. Shouldn’t a small delay be fine?
Not exactly. Here’s why delay kills dictation workflows.
When you dictate and wait for text, your attention shifts. You stop thinking about what to say next and start watching the screen, waiting for text to appear. Fast feedback lets you talk naturally. Slow feedback makes you pause after sentences to check if the text came through correctly. You end up speaking in fragments instead of flowing sentences.
If you need to correct a mistake, an instant tool lets you speak the correction right away. A slow tool forces you to wait, check, then edit — which is slower than just typing.
And these delays add up. If you dictate 20 times a day — writing emails, drafting documents, taking notes — those small pauses compound into real wasted time. Over a year, you lose hours to waiting. See how to dictate into any app on Mac without copy-pasting.
For real-time voice-to-text, speed isn’t a nice-to-have. It’s the difference between a tool that feels good and one that feels like work.
Local vs. cloud: where the delay comes from
To understand why 80ms is exceptional, think about what happens to your voice in different systems.
Cloud-based dictation (Google Docs, Otter, Speechify)
Cloud tools send your voice to the internet and back. Your Mac records your voice, sends the audio to a server somewhere far away, that server converts it to text, and sends the result back to your screen.
All that back-and-forth takes time — typically 225 to 800 milliseconds. No matter how fast your internet is, this round trip has a built-in delay you can’t eliminate.
Hybrid approaches (Apple Dictation)
Apple’s built-in dictation does some work on your Mac first, then sends the rest to Apple’s servers. It’s faster than pure cloud tools, coming in at around 200ms, but it still relies on the internet for the heavy lifting.
Local-only dictation (Dictato)
Everything happens on your Mac. Your voice is recorded, converted to text, and displayed — all without ever leaving your computer.
Total: ~80ms
No internet needed. No servers. No round-trip delay. This is the fastest approach possible, because there’s simply nowhere for the data to travel. That’s also why local speech recognition matters for privacy.
How Dictato achieves 80ms on your Mac
Two things make this possible: modern Mac hardware and smart software design.
Your Mac is built for this
Modern Macs with M-series chips (M1, M2, M3, M4) have a dedicated part of the processor designed specifically to run AI tasks — like converting speech to text. This specialized hardware is much faster at this job than a general-purpose computer would be. Think of it as a dedicated assistant whose only job is understanding your voice, rather than one person trying to do everything at once.
It starts converting as you speak
Most dictation tools wait until you stop talking, then convert your entire recording at once. This is like waiting for someone to finish a whole paragraph before you start writing it down.
Dictato works differently: it starts converting your speech the moment you begin talking, not after you stop. It makes its best guess with the first sounds, then refines as more words arrive. By the time you finish speaking and release the hotkey, the text is already done.
This is why it feels instant — the work happens while you talk, not after.
Three engines to choose from
Dictato includes three speech recognition options. Parakeet is the fastest and most accurate, supporting about 25 languages. Whisper supports 99 languages, with a slight speed trade-off (around 120ms). Apple SpeechAnalyzer (built into macOS 14+) covers 20 languages with zero additional download.
You can pick the one that best fits your language needs and speed preference.
Why most tools are slow
Most dictation tools send your voice to the internet. That round trip is the main reason they feel sluggish.
Cloud tools (Otter, Google Docs, Speechify) rely entirely on remote servers. They need an internet connection and always have a noticeable delay — at least 200-800ms. They also can’t work offline.
Browser-based options add even more overhead on top of the cloud delay, landing at 300-800ms.
DIY local setups (running open-source models yourself) process everything on your Mac, but they use the general-purpose part of your processor instead of the specialized AI hardware. The result: 300-1000ms delays and heavy battery drain.
Apple Dictation is faster than pure cloud, but still sends data to Apple’s servers, adding delay (150-250ms) and raising privacy questions.
The bottom line: cloud tools are slow because of the internet round trip. Most local tools are slow because they don’t take full advantage of your Mac’s hardware. Dictato does both — fully local and fully optimized.
Speed comparison: how Dictato stacks up
Here’s how common dictation tools compare in real-world use:
| Tool | Delay | Works offline? |
|---|---|---|
| Dictato (Parakeet) | ~80ms — feels instant* | Yes |
| Dictato (Whisper) | ~120ms — barely noticeable* | Yes |
| Apple Dictation | ~200ms — slight pause | Only partially |
| Google Docs Voice Typing | 350-500ms — noticeable wait | No |
| Otter | 400-600ms — clear lag | No |
| Speechify | 300-800ms — inconsistent | No |
| DIY Whisper setup | 500-2000ms — significant lag | Yes |
*Note: These times are for pure transcription. Adding optional features like translation or AI proofreading will increase processing time.
The difference between 80ms and 200ms+ is not subtle in practice. It’s the difference between dictation that feels instant and dictation where you’re constantly waiting for text to catch up. For a detailed breakdown, see our Dictato vs Wispr Flow and Dictato vs Superwhisper comparisons.
The real impact on your day
Imagine a lawyer dictating case notes — five sessions a day, two minutes each.
With a slower tool (200ms+ delay), each session includes small pauses: waiting for text to appear, checking it came through right, losing your train of thought. Those micro-interruptions add about 2 extra minutes across the five sessions.
With Dictato (80ms), the text keeps up with your speech. You stay focused. You finish faster.
Two minutes a day doesn’t sound like much. But across a full working year, that’s over 8 hours of recovered time — just from eliminating tiny delays. And if you dictate more frequently (emails, notes, messages), the savings grow.
It’s not just a number on a spec sheet. It’s a tool that stays out of your way.
Why this matters for Mac users
Mac users have historically had limited dictation options. Apple Dictation requires internet for full accuracy. Third-party tools send your voice to the cloud, raising privacy concerns. Many people just accept that voice-to-text “isn’t ready yet” on Mac.
But modern Macs with M-series chips changed the equation. These processors have dedicated AI hardware that can run speech recognition locally, right on your computer, without needing the internet. This wasn’t practical five years ago.
The question is no longer “can a Mac do this?” — it’s “has anyone built software that takes full advantage of it?”
At 80ms, Dictato is the answer.
Putting it together
Real-time dictation at 80ms is not magic. It’s what happens when speech recognition runs entirely on your Mac, using hardware designed for the job and software built to take full advantage of it.
For professionals who dictate regularly — writers, lawyers, developers, researchers — the difference between 80ms and 200ms+ matters more than you’d expect. New to voice typing? Start with our beginner’s guide to dictation on Mac. It’s the difference between a tool that disappears and one you’re constantly aware of.
When your voice-to-text doesn’t feel instant, you stop using it. When it does, it becomes part of how you work.
If you’ve tried dictation before and abandoned it because of lag, the problem wasn’t you. It was the delay. And 80ms fixes it.
Ready to experience instant dictation on your Mac? Dictato brings 80ms voice-to-text to any application, with three speech recognition engines and 100% local processing. Download dicta.to today.