Particle.news
Download on the App Store

MacPaw Partners With Liquid AI to Build On-Device AI Stack for Macs

The deal adapts Liquid AI’s foundation models to run on Apple silicon to keep user data on device.

Overview

  • MacPaw and Liquid AI announced the strategic partnership Wednesday to co-develop a hardware-optimized on-device AI stack that pairs Liquid’s foundation models with MacPaw’s Elix inference and Mnemos memory technologies.
  • MacPaw says its macOS assistant Eney will be the first product to receive the on-device integration later in 2026, with core inference running locally on Apple silicon and cloud models available when needed.
  • The companies are building the stack as shared infrastructure so Mac developers can use it through Setapp, and MacPaw is testing a credit-based pricing model for AI operations in the marketplace.
  • Liquid AI will select and tune model architectures for Apple hardware to increase efficiency while Mnemos will provide persistent on-device conversational memory to retain context across sessions.
  • The partnership reflects a larger industry shift toward on-device AI to improve privacy, cut latency and enable offline workflows, and it could let Mac apps expose local context to assistants for more useful, native automation.