Blog

Apple's Engine or a Local Tier: How to Actually Choose

People ask me which engine they should use, and I used to give a long answer full of nuance. Then I realized the honest answer is mostly a flowchart with three questions, and the nuance is just the footnotes. Let me give you the flowchart, and then the footnotes, because this is a decision you should make in two minutes, not two hours.

Here is the thing to know first. WriteAmp has four ways to make a suggestion. Three are its own local tiers, the small, medium, and larger models I wrote about in the engines post. The fourth is Apple's on-device model on macOS 26, which I wrote about in the trade post. All four keep your words on your Mac. The difference is speed, personalization, and whether you want to download a model. That is the whole decision.

The three questions

If you answered "don't mind the download, want it personal, want it fast," that's the local tier, and for most people the smallest one is the right starting point. If you answered "don't want the download, personalization isn't the point," that's the Apple path. And if you're not sure, the honest default is the smallest local tier, because it's fast, it learns your voice, and it's the version I tuned first.

The tiers, honestly

Let me give you the actual landscape, because "local tier" is three different things and the difference matters.

The bigger local tiers are where people get confused, so let me be direct. The instinct is to pick the biggest, because bigger sounds better. For finishing the next few words of a sentence, that instinct is usually wrong, because speed matters more than raw capability in this window, and I wrote a whole post about why. The bigger tiers exist for people who want the best phrasing and are willing to trade the speed. They aren't the default for a reason.

The honest footnote about Apple's engine

Here is the footnote that trips people up, and it's worth stating clearly. Apple's engine and your personalization are separate by design. Your writing history, the thing that makes suggestions sound like you, is only used by the local tiers. Apple's model doesn't see it, which is good for privacy and bad for personalization, both at once. If you switch to the Apple path, you're trading your voice for the convenience of no download.

And there's one more honest thing about the Apple path that tilts some people toward it. On macOS 26, Apple Intelligence mode stays free even after the trial ends, while the local model is what the license gates. So if you're trying WriteAmp and have not decided whether to buy, the Apple path is the way to keep using the habit while you decide. I wrote about that trade in detail, and it's a real advantage of the bundled engine.

The decision, in one paragraph

Here is the whole guide compressed. If you don't mind a one-time download and you want the fast, personal, tuned version, use the small local tier, and try the bigger ones later if you want better phrasing. If you don't want the download, or you're on a fresh Mac, or you're still in the trial and want to keep the habit going, use the Apple path. Switch any time, feel free, because the engine is plumbing and the habit is the product, and the plumbing should serve the habit, not the other way around.

The suggestion looks the same whichever engine made it, which is the entire point. The trial is free for 30 days if you want to try all four and keep the one that feels right. The benchmarks have the speed numbers, and the engines post has the tier details. The choice is two minutes, not two hours, and now you've the whole map.

---

On macOS 26+ Macs, Apple Intelligence mode stays free even after the trial ends — you always keep a working path to suggestions.

Written by Amit Ashwini, who builds WriteAmp and runs its marketing. More: why the Tab key beats the chat box · mini, midi, and max compared · benchmark methodology.