MacPaw And Liquid AI Build Local Inference Path For SetApp Developers
TechCrunch reported that MacPaw is working with Liquid AI on on-device inference for its Eney assistant and a developer stack for SetApp, shifting part of its AI plan from cloud access to local processing.
MacPaw is building a locally hosted version of its Eney AI assistant with Liquid AI, with the longer-term goal of letting developers run on-device inference through its SetApp app store.
The Ukraine-based app developer unveiled Eney last year.
Under the partnership, Liquid AI will develop Elix, an on-device inference system, along with a local memory system for MacPaw’s assistant.
The arrangement is intended to let assistants and agentic workflows operate offline once the local architecture is ready, according to MacPaw CEO Oleksandr Kosovan.
Liquid AI co-founder and CEO Ramin Hasani said the company chooses a model architecture tailored to the target hardware before training.
That approach, he said, is designed to run intelligence efficiently on the device, with privacy and security benefits.
Liquid AI is also developing a customization stack that would allow models to use user input and improve over time.
Apple already offers local models to developers, but Hasani said Liquid AI’s models are focused on performance across different capabilities.
MacPaw has not said that its system will replace cloud models.
Kosovan said the planned platform would also provide access to models from companies such as Google, combining on-device inference with external providers.
MacPaw plans to put more emphasis on AI apps in SetApp, its subscription-based app store with more than 150,000 paying users.
After finalizing the local processing architecture with Liquid AI, the company wants to make the technology available to developers building their own apps.
The store is also experimenting with credit-based pricing.
Users would be able to perform a certain number of AI operations based on their available credits and the complexity of each task.
