Onhand
FeaturesCompareBlogAboutPrivacySupportTermsLicenses
Menu
FeaturesCompareBlogAboutPrivacySupportTermsLicenses
All documents

About

About Onhand Studio

Onhand Studio runs image generation on your Apple device. This page explains who builds it, how the app is built, how articles are prepared, and how product claims are checked.

Effective and last updated: September 30, 2026

About the author: Mohit P

Mohit P is the founder and developer of Onhand at ArkalEdge. He builds Onhand, an app that generates and edits AI images on the iPhone or iPad itself, so prompts and photos are not uploaded to a generation server. He writes the articles on the Onhand blog, which carry his byline.

Onhand is coming soon to iPhone and iPad and is not on the App Store yet. Corrections to anything he has written are welcome at the address below.

How Onhand is built

Onhand Studio is independently built and reviewed. The app uses Swift, Apple’s MLX framework, and Core ML to generate and edit images on compatible iPhones and iPads.

Articles are reviewed against the app’s implementation and primary sources. A suggested prompt is an example to try; it is not a claim that we generated or tested its result. We distinguish documented features, general guidance, and measurements from a named device and configuration.

Why Onhand runs on the phone

In a cloud image app, your prompt and your photos travel to someone else’s servers, and an image comes back. Whatever the privacy policy says, the data has already been received.

Onhand downloads the model and does the image generation on your device. Prompts and source photos are not uploaded for inference. Catalog updates, downloads, purchases, and limited app analytics use network services; the privacy policy explains them. Images you export or share follow the settings of the destination, including iCloud Photos when enabled.

Performance claims

Generation time depends on the model, quality choice, device, model loading, and temperature. Earlier articles quoted seven and twelve seconds from development. Those figures do not establish a current performance guarantee across the model library, so we have removed them from general product guidance.

A useful benchmark identifies the app build, model package, device, output size, quality choice, and whether loading is included. It also separates a cool first run from sustained use. Until those details accompany a measurement, use the app on your own device to judge the wait.

Corrections and contact

If something here is wrong, please let us know. Email support@arkaledge.com and say which article and which claim. Factual corrections get made and the article’s date updated to reflect the change.

The same address handles questions about the app itself; the support page covers the common ones first.

© 2026 Onhand · by ArkalEdge
FeaturesCompareBlogAboutPrivacySupportTermsLicenses