MacPaw and Liquid AI are developing the Elix on-device inference system for the Eney assistant
MacPaw is working with Liquid AI to develop the Elix on-device inference system for a local version of the Eney assistant. It wants to make the technology available to developers later through the SetApp app store, which is introducing credit-based payments for AI operations.
Ukrainian company MacPaw has partnered with Liquid AI to develop locally hosted AI models. The goal is to create an offline version of the Eney AI assistant, which MacPaw introduced last year. For this purpose, Liquid AI is developing an on-device inference system called Elix and a local memory system.
According to Ramin Hasani, co-founder and CEO of Liquid AI, the company chooses an architecture tailored to specific hardware before training models, which is intended to enable the most efficient version of intelligence running directly on the device, with privacy and security benefits. Liquid AI says it is also building a layer for adapting models, whereby the models are intended to use data from user input to improve and gradually become more intelligent. According to Oleksandr Kosovan, CEO of MacPaw, locally hosted models are intended to allow users to run assistants and agentic tasks offline. Hasani also said that although Apple already offers developers its own local models, the models from Liquid AI focus on performance in different capabilities, according to him.
MacPaw wants to focus more on AI applications in its subscription app store SetApp, which has over 150 000 paying users. Once it completes the local processing architecture with Liquid AI, it wants to make the technology available to developers so they can use on-device inference in their own applications. According to Kosovan, the platform will also offer access to cloud models from other companies, such as Google, serving as a single hub for developers. MacPaw is also already experimenting with a credit-based payment model for the app store, where users perform a certain number of AI operations based on the number of credits and the complexity of the task.
Why it matters
For application developers in the SetApp ecosystem, this promises access to ready-made infrastructure for on-device AI inference and local memory, without having to build their own solution, plus unified access to cloud models from other companies. The credit-based payment model for AI operations also indicates how MacPaw plans to monetize AI features in SetApp. For end users, local processing should mean the ability to use AI assistants and agentic tasks offline with greater privacy, because data does not have to leave the device.
Two audiences, two different impacts
What this means
For individuals
Users of MacPaw applications, including the Eney assistant, should eventually be able to run AI assistants and agentic tasks locally and offline, with privacy benefits, once the local version is available.
More practical updates →For a business
MacPaw is building its own infrastructure for on-device AI inference with Liquid AI and plans to make it available to developers in the SetApp app store alongside a credit-based monetization model. SetApp has over 150 000 paying users — this represents a new product direction and a potential source of revenue.
DevelopmentCheck the original
Event sources
only one source so far · 1 publisher, 1 independent. We count feeds from the same owner only once.