Blog

The Battery Feature Nobody Asks For

Nobody has ever emailed me asking for battery management in an autocomplete app. It is not a feature people request, because it is not a feature people see. It is the thing that happens in the background so that three hours into a flight, your Mac is not begging for the charger because a tiny gray suggestion was busy thinking.

I am going to tell you about it anyway, because it is the feature I am quietly proudest of, and because it explains a lot about how I think about this whole category.

The thing nobody tells you about local AI

Here is the part the marketing never mentions. Running a language model on your laptop is not free. It is not a big deal like rendering a video, but it is real, and it is continuous. Every time you pause and WriteAmp starts thinking about the next few words, the model wakes up and does work, and that work draws power. The bigger the model, the more power.

The honest number is that running one of the larger local engines can cost somewhere in the range of five to eight watts while it is actively generating. That does not sound like much until you are on battery, an hour into a session, and the difference between five watts and eight watts is the difference between landing with twenty percent and landing with six.

The classic fix is to make the user manage it. You pick a small model when you are traveling and a big one at your desk, and you remember to switch. That is what every other tool does, and it is exactly the kind of thing people forget to do, which means they either run the big engine on battery and pay for it, or they leave the small one on forever and never get the good phrasing.

What WriteAmp does instead

I built the switch into the machine. WriteAmp watches whether you are plugged in. The moment you unplug, if you had chosen one of the larger engines, it quietly drops you to the smallest one, the tier that sips power. The moment you plug back in, it restores the engine you actually picked. You do nothing. There is no reminder, no settings screen to visit, no moment where suggestions feel broken.

There is an escape hatch, because I do not believe in features that override people who know what they want. If you are on the larger engine, on battery, and you would rather keep it, there is a pin toggle. Pinning your choice means WriteAmp leaves it alone, and you accept the battery cost. The default is the automatic behavior, because the default should be the thing that is right for almost everyone.

Why the smallest engine is enough on battery

Here is the part that makes the downshift painless. The smallest engine is genuinely good at the job most of the time, because the job is short. Finishing the next few words of a sentence you are already writing does not need a huge model. It needs a fast one. The small tier was built for exactly this, and it is the default for a reason. I wrote about picking between the engines and why bigger is usually the wrong answer, and the battery behavior is that argument in action.

So when you unplug, you are not getting a worse product. You are getting the fast, light version of the product, and the bigger engine comes back the moment you are plugged in again. The only thing you give up on battery is the slightly better phrasing of the larger models, and most people never notice, because they are not A/B testing their own sentences, they are just typing.

The philosophy behind it

The battery downshift is a small feature, but it is the whole design philosophy in miniature. A tool that lives in your typing should take care of the boring things so you do not have to think about them. You should not have to manage your autocomplete's power settings any more than you manage your keyboard's. The tool should be good at its job and then get out of the way, and part of getting out of the way is not draining your laptop while you are trying to work.

That is also why WriteAmp is the way it is about privacy, and performance, and price. I keep coming back to the same instinct: remove the tax, whatever it is. The subscription tax, the context-switch tax, the battery tax. I wrote about the three times WriteAmp uses the network and about what the one-time price actually buys you, and they are all the same story from different angles. I do not want you to have to think about the tool. I want you to type.

The downshift is on by default, so if you download the app and unplug your Mac somewhere, you will not see it happen. That is the point. If you want to watch it work, pick the largest engine, unplug, and glance at the menu bar. The little battery icon is the whole feature saying hello.

The download is free for 30 days if you want to feel it, and the benchmarks are there if you want to see what each engine costs in speed before you trust mine on power.

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.