Typeahead
AI autocomplete for every app on your Mac
Start new thread

Typeahead 2.0 - Private AI autocomplete for every app on your Mac

by
•
Typeahead 2.0 is private AI autocomplete for every app on your Mac, with writing styles per app, support for any language, and smarter context that adapts to how you work. It now feels more customizable and more useful every day, while staying lighter on memory and faster in the background. You also get private Insights to see your time saved. One-time $79 purchase, free updates, no subscription.

Add a comment

Replies

Best
Maker
📌

Hey Product Hunt 👋


Sam here, co-founder of Typeahead.

We launched the first version here a few weeks ago. Since then, we have shipped 8 releases. Today we are back with the biggest update yet.

Typeahead 2.0.

If you are new to it, Typeahead is AI autocomplete for every app on your Mac. As you type, it suggests what comes next right at the cursor. Press Tab to accept.

The autocomplete runs on your Mac, so it works offline and your writing stays on your device.

It is a one-time $79 purchase with no subscription.

What is new in 2.0

  • Writing styles and custom instructions per app. Keep Mail professional and Slack casual, or tell Typeahead to keep things short and friendly in plain English. It switches automatically as you move between apps.

  • Support for any language you write in. Suggestions follow the language you are typing, and the app itself is now localized into 16 languages.

  • Smarter context. Typeahead picks up the conversation you are in so suggestions fit what is actually being said. All on-device.

  • More control over pacing. Choose instant, balanced, or relaxed suggestion timing, and decide exactly where Typeahead shows up. Terminals and code editors are supported too, opt-in per app.

  • Private by default. Typeahead is auto-disabled in password managers, wallets, and finance apps. You control what it can read, app by app.

  • Lighter and faster. The new engine frees up hundreds of MB of memory when you step away.

  • Insights. See how much time you are getting back, computed on your Mac and staying there.


Every one of those 8 releases went to our earliest buyers for free.

That is the whole point of one-time pricing. You buy Typeahead once, and it keeps getting better.

Would love your feedback. I will be here all day answering questions.

Sam

  Many congratulations to Sam and Hiten on shipping the second version of Typeahead. The AppSumo team and I were really impressed with v1, and it’s great to see so many exciting updates in v2. The LTD is a steal deal, hope the community grabs it. :)

Maker

   Thanks Rohan, your support means a lot, great to have you following along from the start!

Buy it once and it keeps getting better, that's exactly what we're going for. Everyone who grabbed V1 just got all of this free, and we have plenty more features on the way 😃

The per-app opt-in for terminals and code editors is the detail I would actually use, since most autocomplete is all-or-nothing. Two things on the local model: roughly what size is it and what is the RAM/battery cost of keeping it resident, and in code-editor mode does it parse the buffer as code (syntax-aware) or just treat surrounding text as plain context? Trying to gauge whether it suggests something useful mid-function or only autocompletes prose.

Maker

 Good questions! We recommend Gemma 4 (around 5GB) and offer a selection of lighter models to suit your Mac. It's easy on memory too, using only a few hundred MB of RAM even with the larger model. It frees up further when you step away and runs only while you're actively typing.

In code editors it works from the surrounding text as context, so it's best for the prose around your code, like comments and markdown. For writing the code itself, we would still recommend a dedicated tool like Copilot.

Got it — so it's a prose/comment layer sitting alongside a code tool like Copilot rather than competing with it. Is syntax-aware buffer parsing something you would consider later, or is staying prose-first a deliberate call to keep the local model at that ~5GB footprint? Just setting expectations before I wire it into my editor.

Curious how Typeahead behaves across different kinds of Mac apps. Since the pitch is AI autocomplete for every app on your Mac, does it work through the standard text fields only, or can it also suggest inside places like design tools, chat apps, terminals, and browser-based editors? The edge cases there would be interesting for dev and creative workflows.

Maker

 Great question. It works anywhere the app has a real text field, native apps, chat apps like Slack and Discord, browsers and web editors, Electron apps included. Smart context capture reads the conversation you're in so suggestions fit what's actually being said, in whatever app you're writing. Terminals and code editors are fully supported too (you can opt in per app in 2.0)!

@samasante This is amazing, especially for fast responses and replies. If it runs on a local LLM, does it also run offline?

Maker

 Thanks Roopesh! Yes, fully offline. The model lives on your Mac, so once it's downloaded you can write on a plane with zero connection. You only need internet to grab the model the first time or pull an update.

Maker

 yes it also runs fully offline!

💡 Bright idea

The per-app writing styles feature is a nice touch. Does it sync those custom instructions if I move to a new Mac, or do they stay local to each device only?

Maker

 Love this idea! Right now your styles and instructions stay local to each Mac, which is by design, everything lives on your device and nothing touches the cloud. We're planning to add an import/export option in the next update so you can carry them over to a new machine. Fully local and account-free, as always ✅

Isn't there a risk that autocomplete for everything, Slack, Mail, terminals, ends up homogenizing how people write? Half the point of switching tone between apps is doing it yourself.

Maker

 Good observation, and it's exactly what we set out to solve. Typeahead knows what you write, how and where, then follows your instructions to match your style, down to dialect, slang and the nuances that make it sound just like you. Slack stays casual, Mail stays sharp, none of it blurs together. Write in whatever voice is yours and Typeahead runs with it 😃

"Insights" that stay on your Mac... so how do I actually see them, carrier pigeon?

Maker

Just open the Insights tab and it's all there, the time you've saved and the keystrokes you've skipped, visualized and cleanly laid out by app. Staying on your Mac just means it's all private, on-device and never sent to the cloud.

Tried couple of local autocomplete tools before and battery drain always the dealbreaker. Curious if Typehead handles that better since it's marketed as lighter now.

Maker

 Should feel much lighter than anything you tried before! And V2's new engine makes suggestions even faster, with a bunch of performance optimizations to make sure it's as energy efficient as possible. It frees up memory when you're not using it, then picks up again the second you start typing.

What’s new in 2.0? I have to try this out now!
Maker

 So much packed into this release! Here are some highlights:

✍️ Writing styles + custom instructions, set per app
🌐 Write in any language (plus the whole app translated into 16 languages)
🧠 More intelligent about which context it pulls from the apps you're using
⏱️ Pacing controls so suggestions match your rhythm
🔒 Private by default, off automatically in password and finance apps
👨‍💻 Opt-in terminals and code editors
⚡ New engine with faster suggestions
🔋 Lighter memory footprint and easier on your battery
📊 Insights showing the time you've saved
✅ One-time purchase, free updates for life

It's live now, would love to hear what you think.

Congrats on 2.0, Sam! The per-app writing style switching is the standout feature for me — most autocomplete tools either annoy you in Slack or sound too casual in client emails. Solving that automatically is a smart move.

Question: with the local model running per-app context, how do you handle apps where the user pastes in a lot of external content (long threads, docs, etc.) — does it adapt on the fly or need a moment to pick up the tone shift?

One-time pricing with 8 free updates already delivered is a strong trust signal too. Rooting for you 🚀

Maker

 Thanks Deepanshu, appreciate it! Your tone stays anchored to the style you've set for that app, so a long pasted thread or doc won't pull your voice off. Context is read live as you type with no warm-up, and it focuses on what's relevant around your cursor rather than the whole paste. So it keeps up on the fly, the style holds your tone and the context sharpens the suggestions.

1234
Next