AI Metadata Generator
Upload an image and get a submission-ready title, description and keyword set.
- Reads the image itself, not the filename
- 40 keywords ordered by relevance
- TXT and per-platform CSV export
6 working tools · 2 need no network at all
The busywork of stock contribution, handled.
Write submission-ready metadata, build prompts for any image model, pull colour palettes, and clean keyword lists. No account, no subscription — bring your own free Gemini key.
- Runs on the Gemini free tier
- Your key stays in your browser
- Colour & keyword tools work offline
The toolkit
Six tools, one workflow
Each does a single job properly, and exports something you can actually submit.
Bulk Metadata Generator
Queue a whole folder and export one CSV ready for bulk upload.
- Up to 100 images per batch
- Rate-limit paced, pausable, retryable
- One CSV for the whole batch
AI Prompt Generator
Turn a plain-language idea into a prompt shaped for the model you actually use.
- Midjourney, Flux, Ideogram, Firefly, ChatGPT
- Per-model syntax and flags
- Recent prompts saved on this device
Image to Prompt
Reverse-engineer a prompt from any reference image.
- Whole image, style only, or subject only
- Written for your target model
- Never names artists or brands
Colour Palette Extractor
Pull the dominant colours out of any image as HEX, RGB and HSL.
- Runs entirely in your browser
- HEX, RGB, HSL and WCAG contrast
- JSON and PNG palette export
Keyword & Filename Optimiser
Clean a keyword list, check agency limits, and build a matching filename.
- Deduplicate, trim, sort, count
- Adobe Stock and Shutterstock limits
- SEO filename generator
How to use
From upload to export in four steps
And the first step is optional for half the toolkit.
Add your key
Paste a free Google AI Studio key into Settings. Only the two AI tools need it — the other two work straight away.
Open SettingsPick a tool
Metadata, prompts, palettes or keywords. Each one is a single focused page with no setup of its own.
See all toolsDrop in your work
Drag an image, paste from the clipboard, or type a description. Format and size are validated before anything runs.
Copy or export
Copy any field to the clipboard, or download a real TXT, CSV, JSON or PNG file — generated in your browser.
Guidelines
What agencies actually accept
The limits every tool here validates against, so a submission is not rejected on a technicality.
Images
- JPG, PNG, WEBP or EPS only, up to 10MB per file.
- Uploads are re-encoded to 1024px on the long edge before analysis — vision models tile at roughly 768px, so more resolution adds latency without improving the result.
- Submit your full-resolution original to the agency; this app only ever reads a copy.
Titles & descriptions
- Keep titles under 200 characters — the cap on both Adobe Stock and Shutterstock.
- Aim for 8–15 words: subject first, then action, then setting.
- No keyword stuffing, no comma-separated lists, and skip filler like “image of” or “stock photo”.
Keywords
- Adobe Stock accepts 49; Shutterstock accepts 50. Generated sets land between 30 and 49, so one list fits either.
- 25–45 is the sweet spot. Fewer gets under-discovered; more dilutes relevance.
- Order matters — buyers and agency ranking both weight the leading terms most, so the strongest keyword goes first.
- Single words only — no phrases, no hyphenated compounds, and no word repeated. Never include trademarks, brand names, celebrities or identifiable private individuals.
Filenames
- Lowercase, hyphen-separated, no spaces or accents, under 80 characters.
- Describe the asset, do not repeat a word already in the name — duplicates add length without search value.
- Avoid names Windows reserves (con, nul, prn, aux, com1–9, lpt1–9); the generator escapes these for you.
Always review AI output before you submit. Gemini describes what it can see, so it can miss context or misread a subject. You are responsible for the accuracy of your metadata and for any release the agency requires.
Ready to start? Get a free Gemini key then add it in Settings.