Bulk CSV Metadata Upload for Stock Sites (Adobe, Shutterstock, Freepik & More)

If you're still typing metadata into a marketplace's web form one image at a time, you're doing hours of avoidable work. Every major stock site accepts a CSV upload that applies titles, keywords and categories to your files in bulk. Here's how it works and how to stop it going wrong.

What a metadata CSV actually is

It's a spreadsheet where each row is one file and each column is a metadata field. The marketplace matches rows to your uploaded images by filename, then fills in the rest. A minimal Adobe Stock row looks like:

Filename,Title,Keywords,Category,Releases
photo-001.jpg,"Woman working on laptop in bright office","woman, laptop, office, remote work, business",Business,

Upload your images, upload the CSV, and every field populates at once.

The catch: every platform's columns are different

This is where contributors lose time. The concept is universal but the exact columns are not:

Adobe Stock — Filename, Title, Keywords, Category, Releases
Shutterstock — Filename, Description, Keywords, Categories, Editorial, Mature content, Illustration
Freepik, Dreamstime, 123RF, Pond5, Vecteezy — each has its own header names, separators and category systems.

Use Adobe's column names in Shutterstock's uploader and it will reject the file or silently misread it. Category values differ too — a category that's valid on one site doesn't exist on another.

Formatting rules that prevent rejected CSVs

Match filenames exactly, including extension and case where the platform is case-sensitive.
Quote fields containing commas (your keyword list) so they aren't split into extra columns.
Use UTF-8 so accented characters and symbols don't corrupt.
Keep keyword counts within each platform's cap (49 for Adobe, 50 for Shutterstock).
Use each platform's real category names, not your own labels.

Get one of these wrong across a 200-row file and you're debugging a spreadsheet instead of shipping.

The multi-platform reality

Most contributors sell the same images on several marketplaces. Doing that by hand means maintaining a different CSV, in a different column layout, with different category mappings, for each site — for every batch. That's the hidden tax of going multi-platform, and it's entirely mechanical work.

Keyword once, export everywhere

The efficient model is to keyword an image once in a neutral form, then generate each marketplace's CSV automatically from that single source. Your judgment goes into the metadata; software handles the tedious per-platform reformatting and category mapping.

That's exactly how StockPhotoTags works: upload a batch, review the AI-generated titles, descriptions, keywords and categories, then download a ready-to-upload CSV in the exact format of Adobe Stock, Shutterstock, Freepik, Dreamstime, Pond5 and more — from the same one batch. No column-name debugging, no per-site retyping. Pair it with a tight batch workflow and multi-platform submission stops being a chore.