Apple introduces Private Cloud Compute (PCC) to handle AI tasks that exceed on-device capability while extending device-level privacy protections to the cloud.

PCC runs on Apple silicon servers with hardware-backed security and hardened software foundations to keep cloud processing aligned with device-level privacy.

PCC can run Apple’s server-based foundation models alongside on-device models, enabling more demanding AI experiences without compromising privacy, and data processed via PCC is not used to train Apple’s models unless the user opts in.

Hedy’s pricing remains unchanged: a free tier with 300 minutes per month, plus Hedy Pro at $12.99 per month or $99.99 per year, with a lifetime license available for $299.

The update reflects a broader trend for Hedy toward multiple processing paths and greater transparency about data handling amid regulatory and policy pressures.

Transcription remains on-device by default for all options, ensuring audio never leaves the device unless a processing option is chosen.

Users now have four AI processing options in Hedy: Local AI (on-device/offline), Cloud AI (via partners in a chosen data region), Apple Private Cloud Compute, and Off (pauses AI while transcription continues).

PCC is designed so personal data used to fulfill a request is not stored or accessible to Apple, with temporary processing and data removal after the response is returned.

PCC differs from traditional cloud AI by reducing data-access components, avoiding general-purpose logging, and excluding remote shells or interactive debugging tools.

Apple Private Cloud Compute provides server-side processing on Apple infrastructure tied to the user’s Apple ID, with data used to fulfill requests not retained afterward, and it’s accessible to third-party apps for the first time since 2024.

Hedy AI introduces Apple Private Cloud Compute as a selectable processing option in Hedy AI meeting assistant with the iOS 27 release (Hedy 3.11).

Hedy emphasizes transparency and user control over processing location, with founder Julian Pscheid highlighting user choice and privacy over policy promises.