All comparisons

Comparison

Onhand vs Conjure

Two private, on-device AI image apps for iPhone. Conjure is already on the App Store and leans on SDXL plus style transfer. Onhand is a narrower studio with curated packs, masked inpainting, and no cloud path, coming soon to iPhone and iPad.

By Onhand · Updated September 18, 2026

Onhand, a private on-device AI image studio for iPhone and iPad

The short answer

Pick Conjure if you want a free app that is already shipping, runs DreamShaper XL Lightning for 1024 by 1024 on-device generation, offers style-transfer photo edits, and advertises zero analytics. Pick Onhand if you want a short curated catalog, strength-controlled image-to-image, masked inpainting on the free model, background removal, and a one-time Pro unlock instead of living entirely inside one SDXL Lightning workflow. Conjure wins on available today and zero-data marketing; Onhand aims at a tighter edit loop once it ships.

Side by side

What to checkOnhandConjure
Where generation runsOn the device GPU, alwaysOn-device via Apple Neural Engine / Core ML
Cloud fallbackNone, by designNone claimed
PlatformsiPhone and iPad, iOS / iPadOS 26 or lateriPhone and iPad; App Store lists iOS / iPadOS 18.6 or later (also Mac / Vision on the listing)
PriceFree download, one-time Pro purchaseFree download as listed 12 September 2026
SubscriptionNoneNone listed for core generation
Account neededNoNo
Primary modelCurated packs (fast SD-class about 2 GB free; Pro adds more, including SDXL Turbo and FLUX.2 Klein 4B)DreamShaper XL Lightning (SDXL), about 1024 by 1024
Photo editingImage-to-image with strength, masked inpainting (free model), background removal, resizeStyle transfer presets / custom prompts; listing says not for precise object removal or exact face preserve
Custom modelsCatalog only: no LoRA, ControlNet, or importWebsite claims bring-your-own Stable Diffusion models
Schedulers / knobsShort menu, style presetsListing: steps, guidance scale, schedulers including DPM++ Karras
App Store privacy labelLimited anonymous analytics (see /privacy)Data Not Collected (as listed 12 September 2026)
App Store statusPre-launch; coming soon (waitlist)Live; 3.3/5 from 22 ratings as listed 12 September 2026

What are these two apps?

Conjure ("Conjure: Offline AI Images") is published by Pixel Management. Its App Store listing and site position it as private Stable Diffusion on Apple silicon: no cloud, no accounts, no tracking. The site names DreamShaper XL Lightning and mixed-bit palettization to fit SDXL on phone memory. Generation and style-transfer editing stay on device.

Onhand is an independently built iPhone and iPad app for people who want to prompt or open a photo and get a result without a settings panel. It ships a curated catalog of downloadable packs, runs generation and editing on the device, and has no cloud switch. It is pre-launch, so there are no public ratings to stack against Conjure yet.

Where does generation happen?

Both apps generate on the device. Conjure's listing and site say prompts and images are not uploaded and that the app works offline after setup. Onhand has the same architectural promise: no generation server, airplane mode after a model pack downloads, no account.

Privacy labels differ. Conjure's App Store privacy section reads Data Not Collected as of 12 September 2026. Onhand uses limited anonymous analytics plus Apple and RevenueCat purchase services; the full list is on the privacy page. If "zero telemetry" is the deciding factor, Conjure's label is the stricter public claim today.

What does each one cost?

Conjure is listed as Free on the App Store on 12 September 2026, with reviews describing core generation as free. Onhand is a free download with a free fast model; Pro is a one-time purchase that unlocks more packs, with no subscription and no image credits. Pro price varies by storefront and is not quoted here while it is still being tested.

If you only need one free SDXL Lightning workflow and the app is already installed from the Store, Conjure is the cheaper path today. Onhand only wins on price shape if you want the deeper packs and prefer paying once for a curated catalog.

How much control do you get?

Conjure exposes more diffusion knobs on the listing (steps, guidance, schedulers) and centers on one strong SDXL Lightning pipeline at 1024 by 1024. Its site also claims custom / bring-your-own Stable Diffusion models.

Onhand deliberately has fewer knobs: pick a pack, prompt, optional style, generate. Packs are curated rather than a free-for-all import surface. There is no LoRA, no ControlNet, and no custom model import.

Honest Onhand limits still apply: negative prompts do nothing on Turbo packs (guidance scale zero); readable text inside images is not promised; which packs run depends on available memory (developer floor: iPhone 14 Pro with 6 GB for the free pack).

Conjure's listing requires 6 GB+ RAM (iPhone 13 Pro and newer class devices; not iPhone 13 / 12 / SE with 4 GB). Its site FAQ mentions iOS 17+, while the App Store compatibility line on 12 September 2026 says iOS 18.6 or later. Trust the Store listing when you install.

What can each one do to a photo you already have?

Conjure leans on style transfer: anime, painting, sketch, comic, pop art, intensity levels, custom prompts. The listing is explicit that this changes overall style and mood and is not built for precise edits like object removal or preserving exact facial features. Some App Store reviews complain that edit mode ignores the source photo. Treat that as user feedback, not a lab test.

Onhand covers a shorter, more surgical list: image-to-image with a strength control on every model, masked inpainting on the free fast model, on-device background removal, plain resize, Refine with history. No outpainting, no AI upscaler, safe-for-work on-device safety check.

If you want "make this photo look like anime," Conjure is aimed at that. If you want "restyle lightly or regenerate this masked region," Onhand is aimed at that.

Which one should you pick?

Pick Onhand if

  • You want strength-controlled restyling and masked inpainting, not only style-transfer presets.
  • You want a curated multi-pack catalog (including non-Lightning looks) behind a one-time Pro unlock.
  • You want an app with no cloud option and a short menu rather than scheduler tuning.
  • You are fine waiting for the launch.

Pick Conjure if

  • You want an app that is on the App Store today.
  • You want free unlimited on-device SDXL Lightning generation with a Data Not Collected privacy label.
  • You care about steps, guidance, and scheduler controls and (per their site) custom models.
  • Style-transfer photo looks are enough and you do not need precise inpainting.

Frequently asked questions

Is Conjure free?

Yes, as listed on the App Store on 12 September 2026 (Free). Reviews describe core generation without paid gates. Always re-check the listing before you buy assumptions.

Does Conjure work offline?

Its listing and site say generation runs on-device and works without internet after setup. Onhand is the same once a model pack has downloaded.

Is Onhand a Conjure alternative?

For people who want private on-device images on iPhone, yes in category. Conjure is the shipped SDXL Lightning plus style-transfer option; Onhand is the curated-pack studio with inpainting and a one-time Pro path, not yet publicly rated.

Which devices are supported?

Conjure's listing: 6 GB+ RAM devices (iPhone 13 Pro class and up; iPad Pro/Air M1+), iOS/iPadOS 18.6+ as of 12 September 2026. Onhand: iOS/iPadOS 26+, memory-gated packs, developer test floor iPhone 14 Pro 6 GB for the free pack. Neither comparison here covers Android.

Can Onhand import custom models like Conjure claims?

No. Onhand only runs its catalog. If bring-your-own checkpoints matter, verify Conjure's current import support in the shipping app. The website claims it; Onhand will not match that.

Is Onhand on the App Store today?

Not yet. Onhand is pre-launch and has no public ratings. Join the launch list on this site to hear first when it ships. Conjure is already available on the App Store.

Sources

Facts about other products come from their own listings and documentation, last checked on the date shown. Corrections are welcome at the address on the About page.