MacPaw is pushing deeper into AI with a plan that centers on local processing, developer access, and a new way for SetApp users to pay for AI tasks. The Ukraine-based app developer has partnered with Liquid AI to support locally hosted AI models in its own products and, later, to offer the same technology stack to developers building for its app store.
The work begins with Eney, MacPaw's AI assistant, and extends toward SetApp, the company's subscription-based app store. SetApp has over 150,000 paying users, and MacPaw is preparing it for AI apps with credit-based plans tied to AI operations.
Why MacPaw Is Building Around Local AI
MacPaw is currently working on Eney, the AI assistant it unveiled last year. The new step is to create a locally hosted version of that assistant, supported by Liquid AI's work on an on-device inference system called Elix and a local memory system.
The core idea is that AI tasks can run directly on a user's device rather than depending only on remote systems. In the source article, Liquid AI's co-founder and CEO Ramin Hasani connected that approach to privacy, security, and hardware-specific efficiency. He said Liquid AI chooses an architecture before training its models, and that the architecture is designed around the hardware.
MacPaw's CEO Oleksandr Kosovan also pointed to offline use as a key benefit. According to him, locally hosted AI models will let users run assistants and agentic workflows offline. That matters for a product like Eney because an assistant becomes more useful when it can operate even when cloud access is limited or unavailable.
The emphasis is not only on putting models on devices. It is also on making those models practical for app workflows. If MacPaw can combine on-device inference with local memory, Eney could become a test case for how the company wants AI to work across its broader software ecosystem.
How Liquid AI Fits Into The Plan
Liquid AI is contributing the technical layer behind MacPaw's local AI ambitions. The partnership covers Elix, the on-device inference system, as well as a local memory system. Together, these pieces are intended to support AI that runs locally and can become part of MacPaw's products.
Apple already provides local models to developers, so MacPaw and Liquid AI are not entering an empty field. Hasani's argument, however, is that Liquid AI's models are focused on performance across different capabilities. The source also says Liquid AI is building a customization stack around its models.
That customization stack is meant to let models use user input and improve over time. In plain terms, MacPaw is looking at AI that is not static. The goal described in the source is for models to become more adaptable and more intelligent as they receive input.
That direction carries an important product implication. A local assistant that can adapt over time could support more personalized workflows while still staying within MacPaw's stated interest in local processing. The source does not describe exactly how that will be implemented, but it makes clear that customization is part of Liquid AI's model strategy.
SetApp Is Being Prepared For AI Apps
The partnership is also about SetApp, not just Eney. MacPaw plans to focus more on AI apps for SetApp, its subscription-based app store. With over 150,000 paying users already on the platform, SetApp gives MacPaw an existing audience for AI software and AI-assisted workflows.
Once MacPaw finalizes the local processing architecture with Liquid AI, it wants to make the technology available to developers. That would let developers use on-device inference inside their own apps built for the app store.
MacPaw also wants the platform to provide access to other cloud models from companies like Google. Kosovan described the intended role as a one-stop shop for developers. That means MacPaw is not framing the future of SetApp AI as local-only. Instead, the source describes a platform that could combine on-device inference with access to cloud models.
For developers, that could simplify how AI capabilities are added to apps. They may not need to treat local models and cloud models as completely separate decisions if SetApp provides a common path to both. The source does not provide technical terms for this developer experience, but the direction is clear: MacPaw wants the app store to become a distribution and tooling layer for AI apps.
Credits Could Shape How Users Pay For AI
MacPaw is already experimenting with credit-based pricing for SetApp. Under that model, users can perform a certain number of AI operations based on the credits they have and the complexity of the task.
This is a notable shift from a simple subscription framing. SetApp is already subscription-based, but AI introduces variable work. Some AI operations may be more demanding than others, and the source says MacPaw is tying credit use to both the number of operations and task complexity.
For users, the clearest effect is that AI access may become something they manage through credits. For developers, credit-based pricing could influence how they design AI features inside SetApp apps. A simple operation and a more complex task may not be treated the same way inside the platform's pricing structure.
The source does not say when this model will become final or how many credits users will receive. What it does show is that MacPaw is preparing SetApp for a market where AI features are not just another app function. They may require their own usage model.
What This Signals For MacPaw
MacPaw's strategy joins several pieces into one plan: a locally hosted Eney assistant, Liquid AI's Elix inference system, local memory, developer access for SetApp, cloud model access from companies like Google, and credit-based pricing for AI operations.
The company's direction is still described as a plan in several places. MacPaw is working on Eney, preparing SetApp for AI apps, and looking to make the local processing architecture available after it is locked in with Liquid AI. That makes this less a finished product announcement than a roadmap for how MacPaw wants AI to sit inside its app ecosystem.
The most important point is that MacPaw is not treating AI as a single feature. It is building around infrastructure, distribution, pricing, and developer access at the same time. If those pieces come together, SetApp could become a platform where local AI, cloud models, and AI app subscriptions are packaged for both developers and users.