Ecommerce Image Dataset – Millions of Tagged Product Photos

The ImageHub ecommerce image dataset is a structured collection of high-quality product photographs gathered from leading retail sources. It covers thousands of SKUs across fashion, furniture, beauty, electronics, and more — making it one of the most comprehensive ecommerce image resources available for AI, visual search, and catalog automation use cases.

What's Inside

Every image in the dataset is associated with a product record containing site_name, source folder, title, and a rich set of metadata tags applied through both folder-based classification and CLIP-based semantic tagging. This dual-tagging approach means images carry both structured labels and high-dimensional embeddings for downstream ML.

• Multi-site coverage — images sourced from Amazon, IKEA, Ashley Furniture, Nike, Anthropologie, Caratlane, Wayfair, and dozens more.
• Primary image flags — each SKU has a designated primary image for clean catalog rendering.
• Verified dimensions & MD5 hashes — every file is fingerprinted for deduplication and integrity checking.
• Folder taxonomy preserved — original source hierarchy is retained so you can reconstruct category trees.

Use Cases

Ecommerce Catalog Automation

Teams building or migrating product catalogs use the ecommerce image dataset to populate listings programmatically. The metadata structure maps directly to common catalog schemas: SKU, category, site, and image sequence number. Import scripts can consume the JSONL metadata files to build product records in Shopify, Magento, or custom headless frontends without manual image mapping.

AI Model Training

Computer vision teams use the dataset to train image classifiers, object detectors, and similarity models. The existing tag structure provides free supervision signal for category, subcategory, and attribute labels. CLIP-derived tags offer semantic labels that generalize better across visual styles than folder-based labels alone.

Visual Merchandising & Creative

Marketing agencies and in-house creative teams use the dataset to source reference images for mockups, A/B tests, and seasonal campaigns. The category organization and consistent primary-image selection mean art directors spend less time culling and more time composing.

Technical Specs

• Format: JPEG/PNG, original resolution preserved
• Metadata: JSONL per source folder + per-image JSON
• Tagging: folder taxonomy + CLIP semantic + ML label proposals
• Deduplication: MD5-based cross-catalog duplicate detection
• Delivery: Secure download link via email within 1–2 business days

How to Request Access

Browse the catalog, filter by category or tag, and submit a bundle request. For large or custom exports — including full-site or combined-category packages — use the custom request form on the catalog page. You'll receive a secure, time-limited download link once your bundle is ready.

Browse by Category

Explore our image dataset organized by product category: