Hedy AI has announced that its cross-platform AI meeting assistant now supports Apple Private Cloud Compute as a user-selectable processing option on iPhone and iPad running iOS 27.
The update, delivered in Hedy 3.11, allows users to choose Apple’s server infrastructure for generating session summaries, detailed notes, chat responses, and live suggestions during meetings.
Apple released iOS 27 on September 14, 2026, marking the first time Private Cloud Compute has been opened to third-party applications since the system originally launched in 2024.
Private Cloud Compute is the server-side extension of Apple Intelligence, running on Apple silicon hardware for requests that are too large to process directly on a user’s device.
Apple states that data sent to Private Cloud Compute is used only to fulfill the request, is not retained once the request completes, and is not accessible to Apple itself.
Apple publishes its server software builds for independent inspection and includes them in the Apple Security Bounty program, allowing security researchers to examine those claims directly.
With the Hedy 3.11 update, users now have four distinct AI processing options to choose from inside the app’s settings menu, each governing how session analysis is handled.
Local AI runs the full analysis pipeline on the user’s own device and is capable of operating entirely offline, while Cloud AI processes data through Hedy’s AI partners within a user-selected data region.
Apple Private Cloud Compute handles processing on Apple’s infrastructure tied to the user’s own Apple ID, and the Off option pauses AI analysis entirely while keeping transcription active.
Hedy discloses within the app that Apple does not publish a data-residency commitment for Private Cloud Compute, and users with regional requirements can continue using Cloud AI with an EU region or the Local AI option instead.
The new option extends a pattern Hedy has built across several recent releases, including Local AI Processing introduced in Hedy 3.2 in May 2026, which moved the entire analysis pipeline onto user hardware.
By default, Hedy transcribes audio on-device, meaning conversation audio never reaches Hedy’s own servers regardless of which AI processing option is selected.
Julian Pscheid, founder of Hedy AI LLC, framed the expanding set of options as a matter of transparency rather than feature differentiation, stating: “Where your conversation gets processed should be your call, not something you find out from a privacy policy.”
Pscheid added that the company’s users include those bound by corporate policy, regulatory requirements, or personal preference, saying “Making it a setting is the only honest way to serve all three, and it is the part of Hedy I care most about getting right.”
On Apple’s approach to server-side privacy, Pscheid said: “Publishing the server builds and inviting researchers to check them is a stronger position than a privacy policy can offer.”
He acknowledged the option’s limitations, noting it “does not replace running the model on your own device, and it does not solve data residency, but it is a real option that did not exist for third-party apps two weeks ago.”
Apple Private Cloud Compute support is available exclusively on devices eligible for Apple Intelligence running iOS 27, and the option does not appear on macOS, Windows, Android, or Hedy’s web application.
Apple applies a daily usage limit that counts against the individual user’s iCloud account rather than against Hedy as the developer, a structure that differs from conventional third-party cloud API arrangements.
Hedy remains free to use with 300 minutes per month, while Hedy Pro is priced at $12.99 per month or $99.99 per year, with a one-time lifetime license available for $299.
