According to TechCrunch, MacPaw is building a local version of its Eney assistant using Liquid AI, complete with on-device inference and local memory. The company plans to open that infrastructure to developers and is preparing Setapp with credit-based pricing for AI app usage.
On-device AI gets sold as a privacy win. The real economics are about cost. Running inference locally lets MacPaw dodge cloud compute bills that balloon with every user request. Once an app store starts metering AI usage, it becomes the economic gatekeeper for agentic software on the desktop.
The question is whether developers actually adopt this. Credit pricing is the signal. If Setapp becomes a real platform for practical AI tools—where speed, offline reliability, and predictable billing matter more than model size—smaller makers get a shortcut to distribution. If it doesn't, it's just another storefront with a meter.
Everyone claims on-device AI is private now. MacPaw is positioning itself as the toll collector for desktop AI where it actually works. Not the flashy stuff. The boring, reliable stuff that runs offline and doesn't crater your cloud bill. If that works, Setapp becomes infrastructure instead of a subscription bundler. If it flops, it's a feature nobody asked for.
Filed to the Technology desk · 6 days ago