MacPaw taps Liquid AI to offer on-device inference to devs building for its app store

MacPaw is taking its AI assistant Eney offline. The company behind the Setapp subscription service is building a local version of the assistant using models from Liquid AI, a move that shifts inference from the cloud to the device itself.

On the surface, this is a product update. Beneath it, the decision is about leverage. MacPaw runs an app store, and app stores live or die by the developers who build for them. By offering on-device inference to those developers, MacPaw is handing them a capability that does not require a server bill or a third-party API key. That is a quiet but real advantage in a market where AI features have become the default ask.

The timing is not accidental. The broader AI market has spent the last two years consolidating around a handful of cloud providers, and developers have grown wary of being locked into someone else's infrastructure. On-device inference is the counter-move. It promises privacy, lower latency, and no per-token cost. For a company like MacPaw, which already positions itself as the user-friendly alternative to the mainstream app ecosystem, the fit is natural.

There is also a strategic layer that has nothing to do with the end user. MacPaw is not just buying models; it is building a moat. Every developer who builds on its local inference stack becomes more invested in the platform. The more tools MacPaw offers, the harder it is for a developer to walk away. That is the quiet machinery of an app store: the value is not in the apps, it is in the gravity.

None of this is a guarantee. Local models still trail cloud models on the hardest tasks, and developers will not abandon the cloud entirely. But the direction is clear. The next phase of the AI platform war is not about who has the biggest data center. It is about who can make the model disappear into the device, and who can convince developers that the infrastructure is already theirs.