Technology

MacPaw's New AI Assistant Will Run on Your Device, Not in the Cloud

Martin HollowayPublished 3d ago4 min readBased on 9 sources
Reading level
MacPaw's New AI Assistant Will Run on Your Device, Not in the Cloud

Ukrainian software company MacPaw has partnered with a company called Liquid AI to power its products with artificial intelligence that runs directly on your computer, rather than over the internet. The deal covers MacPaw's upcoming AI assistant, Eney, and its app store, SetApp, which has over 150,000 paying subscribers.

Most AI assistants today — like Siri or Alexa — send your questions to remote servers, process them there, and send back an answer. MacPaw wants to change that. With Liquid AI's technology, Eney would process your requests on your own Mac, with no internet connection needed. MacPaw says this makes the assistant faster, according to a company newsroom post published in December 2025.

Liquid AI focuses on building AI models that are efficient enough to run on everyday devices. Its co-founder and CEO, Ramin Hasani, said the company designs its models specifically for the hardware they will run on before it even starts training them. That approach differs from the more common method of building large, general-purpose AI models and then trying to shrink them down to fit on a phone or laptop. Hasani also said the models will learn from user data and become smarter over time, though he did not specify how much of that learning happens on the device versus online.

MacPaw CEO Oleksandr Kosovan said the local approach lets users run the assistant and automated tasks entirely offline. Eney, which MacPaw calls a "Computerbeing" — an AI companion designed to think with you, act for you, and adapt to how you work — was first shown in mid-2025. MacPaw has demonstrated it as a Mac assistant you control by voice, one that figures out where your work is happening and helps manage it. Eney and Liquid AI's technology are being woven into MacPaw's system care, cybersecurity, app discovery, and AI assistance features, according to a company newsroom recap published in July 2026.

MacPaw also plans to let outside developers use this local processing technology in their own apps sold through SetApp. The platform will additionally offer access to cloud-based AI from providers like Google, making SetApp a single hub for both local and online AI tools.

The company is testing a credit-based pricing model for SetApp's AI features. Users would get credits to spend on AI tasks, with the cost depending on how complex each task is. The details of how credits translate to specific actions were not shared.

MacPaw is not the only company pursuing on-device AI. Apple, Google, and Microsoft have all built local AI into their operating systems, and free tools have made it easier for developers to run small AI models on ordinary computers. What makes MacPaw's approach different is pairing a curated app store — SetApp's subscription model — with a platform that offers developers both local and cloud AI in one place.

For developers, the appeal is simple: one place to make AI calls, with Liquid AI handling on-device work and cloud providers stepping in when a task is too demanding for the local machine. Whether MacPaw can attract enough AI-focused developers to make this platform matter is an open question. SetApp's 150,000 paying subscribers give it a starting point, but that number is small compared to Apple's App Store or popular standalone AI tools that have reached tens of millions of users.

MacPaw also makes system utilities like CleanMyMac and CleanMyPhone, VPN software called ClearVPN, and cybersecurity tools through its Moonlock Engine division. Eney appears designed to tie these products together — an assistant that can handle system maintenance, flag security concerns, and manage everyday Mac tasks.

In this author's view, the most interesting piece of this story is Liquid AI's broader claim: that designing an AI model for specific hardware from the start can produce better results than taking a large model and shrinking it down. If Eney runs well on Macs without sacrificing quality, it would be early evidence that this approach works for everyday consumers, not just researchers.

MacPaw did not give a specific launch date for the Eney local engine or the developer platform. As of this reporting, the company's newsroom page did not list a dated press release about the Liquid AI partnership.