MacPaw, a Ukraine-based app developer, has partnered with Liquid AI to build on-device AI inference technology called Elix for its AI assistant Eney and SetApp app store.
The on-device approach allows users to run AI tasks offline while maintaining privacy and security.
MacPaw plans to offer this technology to developers on SetApp, alongside cloud models from companies like Google, and is piloting credit-based pricing for users to perform AI operations.
What happened
Ukraine-based app developer MacPaw has partnered with Liquid AI to build Elix, an on-device inference system, and a local memory system to power its AI assistant Eney. MacPaw plans to eventually make this technology available to developers building for its SetApp app store, which has over 150,000 paying users.
Why it matters
On-device AI processing lets users run assistants and agentic workflows offline while keeping data private and secure. MacPaw is now positioning SetApp as a platform where developers can access both on-device inference and cloud models from companies like Google, creating a centralized option for AI app builders.
What to watch
MacPaw is already experimenting with credit-based pricing for the app store, where users earn a certain number of AI operations based on the credits they have and the complexity of the task.
MacPaw, an app developer based in Ukraine, announced a partnership with Liquid AI to bring on-device AI inference to its product ecosystem. The collaboration centers on Elix, an on-device inference system, and a local memory system designed to power MacPaw's AI assistant Eney, which the company unveiled last year. MacPaw now plans to build a locally hosted version of Eney using these technologies. According to Liquid AI's co-founder and CEO Ramin Hasani, the approach begins before model training: "We select an architecture that is different and tailored to the hardware. That allows us to really have the most efficient version of intelligence that runs directly on the device, with benefits like privacy and security." Hasani also highlighted a customization capability: "We are also building a customization stack around models. This means that with user input, the models can use the data and improve. We want our models to be adaptable and become more intelligent over time."
MacPaw's CEO Oleksandr Kosovan noted that locally hosted AI models will enable users to run assistants and agentic workflows offline. While Apple already provides local models to developers, Hasani argues that Liquid AI's focus is on performance for different capabilities. Once MacPaw solidifies the local processing architecture with Liquid AI, the company intends to open the technology to developers building for SetApp, MacPaw's subscription-based app store serving over 150,000 paying users. The platform will act as a one-stop shop, providing access to on-device inference alongside cloud models from companies like Google. To support this ecosystem, MacPaw is already piloting a credit-based pricing model for SetApp in which users perform a certain number of AI operations based on the credits they have and the complexity of the task.
MacPaw's partnership with Liquid AI reflects a broader industry shift toward on-device AI inference as an alternative to cloud-only processing. The company is positioning its SetApp store—which already serves over 150,000 paying users—as a platform that bridges local and cloud AI capabilities. By offering developers both on-device inference through Elix and access to cloud models, MacPaw aims to reduce friction for developers who want to support offline AI functionality without building multiple integrations. Liquid AI's emphasis on hardware-tailored model architectures and customization stacks signals a move away from one-size-fits-all LLM (language model) approaches toward models that can adapt and improve based on user data over time. This matters because it allows developers to offer privacy-preserving AI features while maintaining flexibility for performance-intensive tasks that may benefit from cloud processing.
AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.
Free · takes 30 seconds · unsubscribe anytime
Ask AI anything about this article. Q&As are published on this page for other readers too.
The AI news that matters, in one minute each morning.
Sign up free