App developer MacPaw has partnered with Liquid AI to power locally hosted AI on Macs and to eventually offer the same on-device inference capability to developers through its Setapp marketplace, using models and runtime tuned for Apple hardware locally hosted AI models.
The agreement is a joint R&D effort to combine Liquid AI’s efficient model designs with MacPaw’s Mac-native runtime, centring on an on-device inference engine called Elix and a local memory system to support the company’s assistant, Eney, with the stated benefits of improved privacy, lower latency and offline operation.
MacPaw plans to surface the stack to third-party apps via its subscription app store Setapp, which it is testing with experimental credit-based plans and serves over 150,000 paying users, and Liquid AI co‑founder and CEO Ramin Hasani said, "Before training our models, we select an architecture that is different and tailored to the hardware."
Once the companies lock in the local processing architecture, MacPaw wants to make the inference and memory stack available to developers on Setapp, TechCrunch reported.
