I built DraftKey — on-device AI autocomplete for Mac apps — looking for feedback

Disclosure: I’m the maker of DraftKey (Aura Technologies), so I’m posting this as the developer, not as a neutral user. This isn’t a sale listing; I’m after honest feedback.

What it is: DraftKey is private, on-device AI autocomplete and rewriting for the Mac apps you already use. It suggests completions as you type and can rewrite selected text, without you switching to a separate chat window.

Privacy: The AI runs on-device on your Mac, so your text isn’t sent to a cloud service for the suggestions.

Price (just the facts): $4.99, one-time lifetime license. There is no free trial, and I’d rather say that up front.

Link: https://getdraftkey.com

I’d love feedback from Mac power users:

  • Which apps would you most want this to work in?
  • What would make on-device autocomplete worth using every day?
  • Anything about the privacy or pricing approach that seems off?

Thanks for reading, and I’ll answer any questions here.

2 Likes

My ongoing quest is somewhat related to this but perhaps will be your next project rather than adding to this. In any event - my ideal would be a dictation app that works in both Microsoft Word and web emai which combines dictation and editing i.e. “Put intro as a bold heading… then new paragraph this is a test - no this is 2 tests - start underline and this is an amazing new product. Now clean up the grammar”

So…how would we give you feedback when your app won’t run without a paid license?

3 Likes

I approved the post because it seemed the questions were general enough that actually using the specific app wasn’t necessary.

3 Likes

@webwalrus fair point, and that one’s on me. There’s no trial build right now. The closest thing is the “Try it” section on getdraftkey dot com. To be clear, it plays prepared examples so you can feel the Tab / Cmd+Option flow. It isn’t the live model. Feedback on that flow is still useful to me. So is feedback on the pricing itself ($4.99 once, up to 3 Macs): does “no trial” stop you from trying it at that price?

@rkaplan good use case. DraftKey doesn’t do dictation itself. But anything you dictate with macOS Dictation lands as normal text, and DraftKey works on that text. Select the paragraph, tap Cmd+Option, and pick Fix writing, Clearer, Shorter, or a tone. You review the result before it’s inserted. It works in compatible standard Mac text fields. Some custom editors and secure fields are excluded, so the honest answer for Word and specific webmail editors is “test it in yours”. I’d really like to hear which ones misbehave. Everything runs on-device, so dictated drafts never go to a cloud AI.

I find your claim that you’re here primarily for feedback rather difficult to reconcile with what I’ve found.

This appears to be your very first post in this community, with no prior participation in discussions. And it’s not an isolated introduction.

You’ve posted remarkably similar promotional messages on several other platforms, including:

Your Hacking with Swift account was created on October 28, and its only two visible posts promote your own commercial applications, DraftKey and TokenBar.

There are also numerous articles on DEV.to and Medium promoting DraftKey, including comparisons with competing products.

Now, there’s nothing inherently wrong with marketing your software. But presenting what appears to be a coordinated, cross-platform promotional campaign as a genuine attempt to engage with individual communities strikes me as rather disingenuous.

There’s also a separate issue I’d like you to address: transparency.

Your application requires powerful macOS Accessibility permissions and processes text from other applications. Yet users have no practical way to independently verify your privacy and security claims from the information currently available.

So I’d appreciate answers to a few questions:

  1. Was DraftKey developed using AI coding tools? If so, to what extent?
  2. Has the application undergone any independent security review, particularly regarding its Accessibility access and handling of potentially sensitive text?
  3. Beyond publishing promotional material, how do you intend to contribute to this community?

I’m not suggesting your software is malicious or that using AI to develop applications is inherently wrong.

But there’s a difference between participating in a community and using it as another distribution channel for your marketing campaign. And I think that distinction deserves an honest discussion.

3 Likes

Being completely transparent, it’s one of many signals.

Negative signals:

  • This is an app being posted by a developer, not a community recommendation.
  • That developer just joined the community, for the purpose of posting the app.
  • The app has no trial.
  • I haven’t checked, but I’m guessing when I download the app it’s going to want to monitor input and/or have access to my files - which is a lot of access.
  • The price is $5 for a lifetime license. It seems weird for that to be a negative, but at $5 for a niche app it suggests you don’t have a business model for continued development. And anything that gets into the weeds with keyboard input and on-device AI will need future development.
  • This looks to be your first app, as Aura Technologies has a website which is (a) sparsely populated, and (b) doesn’t mention any apps at all.

Positive signals:

  • Checkout is through a reputable third party.
  • You responded to my question. I didn’t expect that, and it’s appreciated.

That’s what goes into the proverbial blender when I’m considering a purchase. I’d have all the same questions @johnkree has as well, particularly given that the Aura website mentions “We’re building what comes next, using AI to turn ambitious ideas into everyday possibilities,” which sounds like vibe-coding (as distinct from AI-assisted coding).

But the above is my direct observation.

2 Likes

Lack of a trial period and $5 pricing are the biggest red flags.

Assuming 100% genuine intent and motivation to succeed in software development, these two together suggest almost 0% likelihood that your app will survive.

Almost nobody will buy an app that does not have some sort of trial option.

And $5 lifetime pricing is silly. Either make it open source because it is a labor of love and you seek help from others, or price it realistically so you can survive and support it long-term.