Learning how to write metadata for stock photos is the single highest-leverage skill a contributor can develop. Your photo can be technically flawless and beautifully composed, but if its title, description, and keywords don't match what buyers type into the search bar, it will never be seen — and an unseen photo earns nothing. Search is how more than 90% of stock content gets discovered, and search runs entirely on metadata.
The good news: metadata is a system, not an art. Agencies like Adobe Stock, Shutterstock, and Getty Images all reward the same fundamentals — accurate titles, relevant keywords in the right order, and clean, honest descriptions. Once you understand those fundamentals, you can apply them to every file you upload, manually or with automation.
This guide walks through every layer of stock photo metadata in 2026: titles, descriptions, keyword selection and ordering, keyword count, IPTC embedding, common rejection traps, and the workflow that lets top contributors tag hundreds of images per day.
Why Metadata Decides Whether Your Photos Sell
Every stock agency is a search engine with a checkout button. When a designer needs an image of a "diverse team meeting in a modern office," the agency's algorithm scans titles, keywords, and descriptions to build the results page. Files whose metadata closely matches the query — especially in the title and the first keywords — rank higher. Files that rank higher get more views, more views produce more downloads, and downloads feed back into the ranking algorithm. Metadata is the entry ticket to that flywheel.
Metadata also affects acceptance. Reviewers at Adobe Stock and Shutterstock check whether your title and keywords accurately describe the image. Irrelevant keywords ("keyword spam") are one of the most common rejection reasons across agencies, and repeated violations can hurt your account standing. Well-written metadata therefore does double duty: it gets files approved faster and sells them once they're live.
Stock Photo Keywording Tips: Start With the Title
The most useful of all stock photo keywording tips is this: write the title first, and write it like a buyer's search query. A strong title is a natural, descriptive sentence fragment of roughly 5 to 15 words that states the subject, the action, and the setting. Compare these two titles for the same file:
Weak: "IMG_2871 edited final" — or even "Business meeting." Strong: "Diverse business team collaborating around a laptop in a bright modern office." The strong version contains the subject (diverse business team), the action (collaborating around a laptop), and the setting (bright modern office) — three separate search phrases a buyer might use, packed into one readable sentence.
Practical title rules that hold across agencies in 2026: keep it under about 200 characters (Adobe Stock truncates around 200; Shutterstock descriptions can run to 200 as well), avoid camera jargon and file names, don't stuff synonyms, capitalize normally, and never include trademarked brand names for commercial content. If the image is AI-generated, most agencies now require that fact to be indicated during submission — but keep the title itself descriptive, not "AI image of..."
Descriptions matter slightly less than titles for ranking, but where a separate description field exists, use it to add context the title couldn't fit: mood, concept, use case, season, or copy space. One or two natural sentences is enough.
How Many Keywords for Stock Photos? The Real Answer
Contributors constantly ask how many keywords for stock photos is optimal. The technical limits are similar everywhere — Adobe Stock and Shutterstock both cap submissions at 50 keywords, iStock at 50 through DeepMeta — but the practical answer is not "always use all 50."
A useful target is 25 to 45 genuinely relevant keywords. Every keyword should pass a simple test: if a buyer searched this exact word, would my image be a satisfying result? If the answer is no, the keyword is noise. Noise keywords don't just risk rejection — they actively hurt you, because downloads-per-view is a ranking signal. When your photo appears for irrelevant searches and gets skipped, agencies learn to rank it lower everywhere.
Order matters more than count. Adobe Stock explicitly weights the first ten keywords most heavily, and reviewers read them first. Front-load your keyword list in this order: (1) main subject, (2) action or concept, (3) setting and secondary subjects, (4) style, mood, and composition terms like "copy space" or "top view," (5) broader category terms. A landscape shot of a mountain sunrise should start with "mountain, sunrise, landscape, nature, summit" — not with "beautiful, amazing, wallpaper."
Best Keywords for Stock Photos: Think Like a Buyer, Not a Photographer
The best keywords for stock photos describe what the image is for, not just what it is of. Buyers are designers, marketers, and editors searching for images to communicate an idea. A photo of a person staring at a laptop at midnight is literally "man, laptop, desk, night" — but it sells on concept keywords like "deadline," "overwork," "burnout," "remote work," and "stress." Always tag three layers:
Literal layer — who and what is visible: "woman, laptop, kitchen table, coffee, notebook." Conceptual layer — what the image communicates: "freelance, work from home, productivity, small business, work-life balance." Technical/stylistic layer — how it's shot and how it can be used: "candid, natural light, copy space, horizontal, close-up." Files tagged across all three layers surface in dramatically more searches than files tagged only literally.
Where do you find high-demand terms? Type your subject into the agency's own search bar and study the autocomplete suggestions — those are real buyer queries. Check the keywords on best-selling images in your niche. And describe people accurately but respectfully: age group, ethnicity if relevant to search ("asian woman," "senior man"), and the number of people, since buyers filter by all of these.
IPTC Metadata for Stock Photography: Write Once, Upload Everywhere
IPTC metadata for stock photography is the industry-standard way to embed your title, description, and keywords directly inside the image file itself. When you fill in the IPTC "Title," "Description," and "Keywords" fields — in Lightroom, Photoshop, Bridge, or an automated tool — that information travels with the JPEG. Adobe Stock, Shutterstock, Dreamstime, and most other agencies read embedded IPTC on upload and pre-fill the submission form automatically.
This is the key to multi-agency efficiency: keyword a file once, embed the metadata, then upload the same file to five agencies without retyping anything. Combined with FTP batch uploading, embedded IPTC turns a full evening of form-filling into a few minutes of drag-and-drop. Note the difference from EXIF: EXIF stores camera data (shutter speed, ISO) and is ignored for search; IPTC stores descriptive data and is exactly what agencies index.
Using an AI Metadata Generator for Stock Photos
Manual keywording takes 3–5 minutes per image when done properly. At 100 images, that's an entire working day spent typing — which is why an AI metadata generator for stock photos has become standard kit for serious contributors. Modern vision models analyze the actual content of your image and generate a descriptive title, a natural description, and a ranked keyword list in seconds, covering the literal, conceptual, and stylistic layers automatically.
Rastock AI was built specifically for this workflow: it generates agency-ready titles, descriptions, and keywords tuned to Adobe Stock, Shutterstock, and Getty requirements, embeds them as IPTC metadata, and can then push your files to the agencies via built-in FTP bulk upload. You can review and edit every field before anything is submitted — the AI does the heavy lifting, you keep editorial control. See the full feature breakdown on the Rastock features page.
The economics are hard to argue with: if your time is worth even $15/hour, manually keywording 500 images costs you $375–$625 in labor. Automated tagging plus a quick human review cuts that by roughly 90%, and Rastock's pricing starts free, so you can test the output quality on your own portfolio before committing to anything.
Common Metadata Mistakes That Trigger Rejections
Keyword spam is the biggest offender: adding trending terms ("christmas," "business," "ai") to images that have nothing to do with them. Reviewers catch it, and agencies penalize it. Close behind: trademarked terms (brand names, product names, "iPhone," "Nike") in commercial submissions; naming recognizable people or claiming a location you can't verify; using "editorial-style" captions on commercial files; duplicated titles pasted across visibly different images; and leaving auto-generated camera filenames as titles. Each of these is an easy rejection that costs you review-queue time.
A subtler mistake is inconsistent language: agencies expect metadata in English (they auto-translate for international buyers). Mixed-language keyword lists confuse both reviewers and search algorithms. Write everything in clear, simple English and let the platforms handle localization.
A Repeatable Metadata Workflow for 2026
Here is the workflow top contributors converge on. First, cull and edit your shoot, exporting final JPEGs. Second, run the batch through an AI metadata generator to produce titles, descriptions, and 30–45 keywords per file. Third, spend a few seconds per image reviewing: fix any misidentified subject, reorder the first ten keywords, add niche terms the model couldn't know (event names, specific species, local context). Fourth, embed everything as IPTC and upload once via FTP to all your agencies. Fifth, track which keywords appear on your best-selling files and feed those learnings back into future shoots.
The same principles extend beyond the big agencies — if you also publish design assets, our guide to Canva contributor tagging strategy shows how to adapt keyword thinking to template marketplaces.
Metadata isn't glamorous, but it compounds. Every properly tagged file keeps working for you in search results for years. Master the system once, automate the repetitive parts, and your portfolio's earning ceiling rises with every upload.
Related reading
Best Keywords for Stock Photos: Data-Backed Strategies
Image Keywords for SEO: The Complete 2026 Guide
Keywords for Photography: What Actually Works in 2026
Frequently asked questions
What metadata do stock photos need?
Three fields drive stock photo search: a descriptive title (5–15 words stating subject, action, and setting), a keyword list of 25–45 relevant terms ordered by importance, and, where supported, a short natural description. Embedding these as IPTC metadata lets agencies auto-fill them on upload.
How many keywords should I use for stock photos?
Adobe Stock and Shutterstock allow up to 50 keywords, but 25–45 genuinely relevant terms is the practical sweet spot. Irrelevant filler keywords reduce your downloads-per-view ratio, which lowers your search ranking over time. Always put the 10 most important keywords first.
Do keywords really affect stock photo sales?
Yes — search is how the vast majority of stock content is found, and agency search engines rank results primarily on how well titles and keywords match the buyer's query, combined with the file's download history. Better keywords mean more impressions, and relevant impressions convert into sales.
What is IPTC metadata and why does it matter for stock photography?
IPTC is the industry-standard format for embedding title, description, and keywords inside an image file. Most agencies, including Adobe Stock and Shutterstock, read IPTC on upload and pre-fill your submission form — so you can keyword once and upload the same file everywhere without retyping.
Can AI write metadata for stock photos?
Yes. AI metadata generators like Rastock AI analyze the image content and produce agency-ready titles, descriptions, and ranked keywords in seconds, then embed them as IPTC and even upload via FTP. Best practice is a quick human review pass to fix niche subjects and reorder the top keywords.
Why was my stock photo rejected for metadata?
The most common metadata rejections are keyword spam (irrelevant or trending terms that don't match the image), trademarked brand names in commercial submissions, camera filenames used as titles, and titles that don't accurately describe the content. Fix the metadata and resubmit — the image itself is usually fine.