MacPaw partners with Liquid AI to bring private, on-device AI to Mac developers

MacPaw, a developer of productivity software for macOS, has joined forces with Liquid AI to embed locally-run AI capabilities into its products and extend them to third-party developers through its app marketplace.

The partnership aims to give Mac users control over their data by processing AI tasks directly on their devices, rather than sending information to cloud servers. MacPaw is developing a local inference system called Elix, along with a on-device memory framework, to power an expanding suite of AI-driven features.

The Local AI Strategy#

MacPaw is building a private version of Eney, its AI assistant launched last year, that will operate entirely on-device. By working with Liquid AI—a startup focused on designing efficient machine-learning models—the company hopes to run sophisticated AI without sacrificing speed or relying on internet connectivity.

Liquid AI's approach differs from typical pre-built models: the startup tailors its architecture to specific hardware capabilities before training begins. "That allows us to really have the most efficient version of intelligence that runs directly on the device, with benefits like privacy and security," according to Liquid AI co-founder Ramin Hasani.

MacPaw CEO Oleksandr Kosovan highlighted another advantage: users will be able to run AI assistants and automated workflows without an internet connection.

Developer Platform and Pricing#

Beyond its own products, MacPaw intends to open the infrastructure to developers using SetApp, its subscription app store serving over 150,000 paying subscribers. Developers will gain access to on-device inference alongside optional cloud models from providers like Google, creating a unified platform for AI deployment.

The company is already testing a credit-based model, where users accumulate credits tied to AI operation complexity. This approach lets SetApp users budget their AI consumption without paying per-query or per-transaction.

The Competitive Landscape#

Apple offers local AI models to its developer community, but Liquid AI contends its models emphasize adaptive performance across diverse use cases. The startup is also building customization layers that let models learn from user data over time, potentially improving accuracy and relevance as they operate.