MacPaw Partners with Liquid AI for On-Device Inference Solutions for App Store Developers
MacPaw, an app development company based in Ukraine, has teamed up with Liquid AI to enhance its offerings with locally hosted AI technology. The collaboration aims to eventually provide developers with access to this tech stack. Additionally, MacPaw is gearing up to launch its app store, SetApp, with credit-based plans specifically for AI applications.
The firm is currently advancing its AI assistant, Eney, which was introduced last year. Plans are underway to create a locally hosted variant, leveraging Liquid AI’s expertise to build an on-device inference system named Elix along with a local memory framework.
Liquid AI’s co-founder and CEO, Ramin Hasani, emphasized the importance of selecting unique architectures suited to specific hardware before training their models. This strategy aims to create the most efficient intelligence that operates directly on user devices, offering enhanced privacy and security features.
MacPaw’s CEO, Oleksandr Kosovan, highlighted that local AI models will empower users to operate assistants and workflows even without internet connectivity.
WhileApple offers its localized models to developers, Hasani points out that Liquid AI focuses on optimizing performance for various functions.
“We are also developing a customization framework around our models. This allows them to learn and evolve with user input. Our goal is to make our models adaptable and increasingly intelligent over time,” Hasani noted.
MacPaw is intensifying its focus on AI applications for its subscription-based app store, SetApp, which boasts over 150,000 paying subscribers.
After establishing the local processing infrastructure with Liquid AI, MacPaw plans to offer this technology to developers, enabling them to utilize on-device inference in their applications. According to Kosovan, the platform will also serve as a gateway to other cloud-based models from companies such as Google, allowing for a comprehensive resource for developers.
Moreover, MacPaw is exploring credit-based pricing for its app store, where users will be able to execute a set number of AI tasks depending on their credit balance and the task’s complexity.
