Adobe Stock AI Image Requirements: Resolution, Upscaling and Submission

Can AI-generated images be sold on Adobe Stock?

Yes. Adobe Stock accepts eligible generative-AI photos, illustrations, vectors, and videos. It requires contributors to review the terms of their generation provider to ensure the content can be licensed commercially and to follow Adobe's restrictions concerning intellectual property, references to real people, disclosure and quality.

When the content was created using generative AI, select Created using generative AI tools in Adobe Contributor. If the content portrays fictional people or property, the applicable People and Property are fictional declaration may also be required. The conditions for content that only received limited AI-assisted editing can differ. Adobe's generative-AI guidelines.

What image resolution does Adobe Stock require?

For photos, Adobe specifies a minimum of 4 MP and a maximum of 100 MP. Divide width × height by 1,000,000 to calculate megapixels; do not confuse the “1K,” “2K,” or “4K” name of an AI generation setting with verified output dimensions.

These calculations are reproducible. The table does not mean that an image meeting the minimum will pass Adobe's visual, copyright, release, or AI-disclosure checks. Nor does it establish that all Adobe content types use identical requirements.

What is the difference between resizing and AI upscaling?

Traditional resizing creates additional pixels by interpolating from existing ones. It may increase dimensions without restoring missing fine detail. Some AI upscaling techniques attempt to infer textures or reconstruct edges, but they can also add false detail, sharpen artifacts or alter anatomy.

For Adobe Stock, the file must be useful at full resolution. A picture with distorted fingers, strange reflections or artificial-looking textures will not become acceptable merely because its width was multiplied by four. Adobe explicitly cautions against increasing resolution in ways that degrade image quality. Adobe AI FAQ.

When does upscaling make sense?

Use it when an otherwise promising image is below the minimum dimension requirement, and the source contains enough coherent detail to benefit from enlargement. If the image is already large enough and clean at 100% magnification, extra processing may be unnecessary. If it has severe generation errors, rerendering or targeted editing may be the better option.

The visual example included with this editorial package uses a synthetic illustration exported as two JPEG files with measurable dimensions. It demonstrates how the megapixel threshold works. It is not an example of genuine generative-AI super-resolution and must not be presented as one.

Adobe Stock photo specifications beyond resolution

For photo submissions, Adobe also requires JPEG format, the sRGB color profile, a 45 MB maximum file size, sharp focus, suitable lighting, and the absence of watermarks or embedded commercial marks. Check source and output formats directly: a file with a .jpg extension is not necessarily a valid JPEG if the export pipeline handled it incorrectly.

Inspect the final saved file—not just the browser preview—for its dimensions, MIME/format, color profile, filesize and obvious visual flaws. Adobe photo requirements, updated September 2026.

A practical visual QA checklist at 100% zoom

Adobe considers image quality and distinctiveness, not just file specifications. An approved-looking preview should still be inspected at the final upload size. Adobe generative-AI submission guidelines · Distinct submission guidance.

• Hands and faces: fingers, teeth, eyes, skin texture and anatomy should be plausible.
• Objects and reflections: watch for warped glassware, impossible geometry and inconsistent reflections.
• Lighting: check shadows, highlights and directional consistency.
• Text and trademarks: look for garbled signs, accidental logos, branding and protected content.
• Upscaling artifacts: watch for repeated patterns, halos, fabricated textures and oversharpening.
• Portfolio distinctiveness: avoid submitting near-identical images with only minor crops or cosmetic variations.

Should an AI image be submitted as a photo or illustration?

Adobe advises submitting photorealistic images that could plausibly have been captured by a camera as Photos. Other generative-AI images that are more illustrative, stylized, or fantastical generally belong under Illustrations. The visual characteristics—not merely the fact that an AI tool created the asset—inform the choice.

Rights and disclosure requirements still apply whichever type you choose. Avoid representing fictional events as documentary journalism or referencing protected people and properties without the necessary permissions.

What should Microstock Studio users do today?

Microstock Studio currently supports AI generation, human image review, and AI-assisted metadata preparation. Integrated AI upscaling is still a planned pre-beta improvement, not a feature we are claiming is released. Until that stage is available and tested, creators should check every output's true dimensions, format, color profile and visual quality before using their existing export workflow.

For the title and keyword side of the submission, see Adobe Stock Keywords. For bulk metadata files, read Adobe Stock CSV Upload Guide.

Frequently asked questions

Key factual answers on Adobe Stock AI image acceptance, resolution thresholds, and labeling standards based on current official guidelines.

• Adobe — Technical and legal requirements for photos
• Adobe — Generative-AI content guidelines
• Adobe — Generative AI FAQ
• Adobe — Distinct generative-AI submissions