Generate AI images with Midjourney and sell them as stock photos on Adobe Stock

The difference between images that sell and images that sit unseen is commercial intent. Buyers on Adobe Stock are designers, marketers, and creative professionals — they are looking for images to put in presentations, advertisements, websites, and reports. They are not looking for artistic expression.

Categories with consistently high download demand: - Business & technology: diverse professionals in office settings, team collaboration, person using laptop, video call scenarios, AI and data visualisation concepts - Healthcare & wellness: doctor-patient interactions, healthy lifestyle, mental health concepts, medical equipment - Sustainability & environment: green energy concepts, recycling, eco-friendly themes, climate and nature - Diverse people: authentic representation of various ethnicities, ages, and body types in everyday scenarios - Concept imagery: abstract visualisations of growth, success, connection, innovation (used heavily in editorial and corporate)

Categories with lower AI approval rates (avoid initially): - Real-world locations (architecture, cityscapes) — AI accuracy on specific real places is poor - Products with brand or text (AI generates hallucinated text and logos) - News and editorial (very specific standards)

Ask Claude: *"I'm generating AI images for Adobe Stock. List 20 specific commercial stock photo concepts in the [business/healthcare/sustainability] category that are in high demand but not completely saturated. Include the key visual elements buyers typically search for in each."*

Log in at [midjourney.com](https://midjourney.com) and begin generating your first batch. The goal is 50 high-quality images — expect to generate 150–200 variations to select the best 50.

Prompt formula for commercial stock photography: ``` [specific scene description], professional stock photography, [lighting description], clean background, [composition], commercial use, --ar 3:2 --style raw ```

Example prompts that produce commercially usable output: - `two diverse professionals reviewing data on a laptop together, professional stock photography, natural office lighting, shallow depth of field, warm and collaborative mood --ar 3:2 --style raw` - `woman meditating in a bright minimalist room, wellness lifestyle stock photo, morning light, calm and serene, simple background --ar 3:2 --style raw` - `abstract concept of artificial intelligence network connections, blue and white, clean corporate style, technology background --ar 16:9 --style raw`

Aspect ratios to generate: - `--ar 3:2` (landscape, most versatile for web and print) - `--ar 1:1` (square, used heavily for social media templates) - `--ar 4:5` (portrait, strong for vertical social formats)

Selection criteria: only keep images where (a) the subject is clearly identifiable, (b) hands and faces look natural — this is the most common AI failure point, (c) there is no hallucinated text, logos, or brand marks, and (d) the composition feels intentional rather than accidental.

Be strict at the selection stage. A portfolio of 50 strong images is significantly more valuable than 100 mediocre ones.

Discoverability on Adobe Stock is driven almost entirely by keywords and titles — buyers search, not browse. A strong image with poor metadata will not sell. A mediocre image with excellent keywords will consistently outperform it.

Title format: descriptive, keyword-rich, 5–15 words. Lead with the subject: "Diverse Business Team Collaborating on Laptop in Modern Office" not "Team Photo".

Keywords: Adobe Stock accepts 25–50 keywords per image. Fill them all. Use a mix of: - Specific: what's literally in the image ("laptop", "woman", "office", "blue shirt") - Conceptual: what the image communicates ("collaboration", "teamwork", "technology", "innovation") - Emotional: the mood or tone ("professional", "friendly", "modern", "approachable") - Use-case: where a buyer would use it ("website header", "presentation background", "blog illustration")

Ask Claude to generate keywords in bulk: *"I have a stock photo showing [describe the image in detail]. Write 40 keywords that buyers searching for this type of image would use — mix specific visual elements, concepts, emotions, and use cases. Output as a comma-separated list."*

Do this for each image, or batch similar images together with a shared keyword set plus 5–10 unique tags per image.

Adobe's category selection: choose the most accurate primary category — this affects where the image appears in category browsing.

Upload images through the contributor portal at [contributor.stock.adobe.com](https://contributor.stock.adobe.com). You can upload up to 100 images at once via web browser or via FTP for larger batches.

During upload, for each image: 1. Select content type: Generative AI (mandatory for AI-generated images) 2. Enter your title 3. Add keywords (paste from your Claude-generated list) 4. Select the primary category 5. Set release requirements: unless your image includes recognisable people or private property, no model or property release is required for AI-generated conceptual imagery

Submit and wait: Adobe's review team evaluates submissions for quality, commercial utility, and adherence to content guidelines. First-time contributors typically wait 3–5 business days for initial reviews. Once established, reviews often happen within 1–2 days.

Understanding rejections: Adobe sends rejection reasons. Common ones for AI images: - Quality issues: soft focus, noise, unnatural-looking hands or faces — rework the prompt and regenerate - Commercial utility: image lacks clear commercial application — focus on clearly purposeful subject matter - Technical issues: resolution too low, wrong colour profile — save at maximum quality and in sRGB

Keep a spreadsheet tracking which images were approved vs. rejected and why — this is your feedback loop for improving future batches.

Stock image income compounds slowly — the first meaningful passive income typically appears 2–4 months after your first upload batch. Here is how to build efficiently:

Weekly upload target: 20–30 new images per week. At this pace, you'll have 200 accepted images within 3 months — the portfolio threshold where consistent monthly income typically begins.

Reading your analytics: Adobe Stock Contributor provides a dashboard showing your image-level performance — which images were downloaded, how many times, and what they earned. After 60 days, review your top 20 downloads and your 20 images with zero downloads. Generate more of what's selling; stop generating what isn't.

Seasonal planning: stock image demand follows a predictable seasonal calendar. Start generating Christmas images in September. Valentine's Day content in December. Summer lifestyle in March. Submit 6–8 weeks before the holiday — Adobe review takes time, and buyers search early.

Diversify across categories: once you have 50 accepted images in your first category, expand to a second. Diversity reduces the risk of any one category's demand drying up and gives you more search entry points across the Adobe Stock catalogue.