Bulk Image Metadata Generator

What Is the Bulk Image Metadata Generator?

The Bulk Image Metadata Generator uses AI vision models to automatically create SEO-ready titles, descriptions, and keyword tags for a whole folder of JPG images in one go.

Just select a local folder, pick which fields you want, and the tool analyzes every image in your browser and hands back clean, editable metadata. You can export the result as CSV or JSON, copy individual rows, or re-embed the metadata directly into the JPG files (EXIF) and download them as a ZIP.

Everything runs client-side through Puter.js — your images never leave your machine for a third-party server you don't control, and you don't need any API keys.

How to Generate Metadata for a Folder of Images

1. Click "Choose folder" (or drag a folder in) to load all JPG/JPEG files from a local directory.

2. Choose which fields to generate — title, description, tags — and configure model, tag count, target language, and style preset (SEO, stock photography, e-commerce, social media, etc.) or write a fully custom prompt.

3. Press Generate. Each image is sent to the selected vision AI model through Puter.js and the results populate an editable table.

4. Edit any row if you want to fine-tune the output, then export: CSV, JSON, per-row copy, or embed-into-JPG + ZIP download.

When Bulk AI Metadata Saves You Hours

Stock photographers uploading to Shutterstock, Adobe Stock, or Getty need titles, descriptions, and keyword lists for every submission — this tool handles hundreds of images in a single batch.

E-commerce stores with large product catalogs need unique, SEO-optimized titles and alt text for every product image. Bulk generation eliminates the bottleneck.

Bloggers, journalists, and content teams can tag large image libraries for search and DAM systems without manual cataloging.

Digital asset managers and archivists can enrich image collections with searchable descriptions and keywords in minutes instead of days.

Tips for Better AI-Generated Image Metadata

Pick the right model for the job. Lightweight models (gpt-4o-mini, gemini flash) are fast and cheap for straightforward subjects. Frontier models (gpt-4o, claude sonnet) catch more nuance for abstract, artistic, or complex scenes.

Use a style preset that matches your destination. "Stock photo" produces keyword-rich, search-optimized output. "E-commerce product" focuses on specs and benefits. "Social media" leans punchy and hashtag-friendly.

Set a sensible tag count — 15-25 is the sweet spot for stock platforms, 5-10 for blog alt-text.

Always review and lightly edit the output. AI is a strong first draft, not a final proof.

Embed the metadata into the JPG EXIF if you plan to upload — most stock and DAM platforms read EXIF ImageDescription, XPTitle, and XPKeywords automatically.

Privacy and How Puter.js Works

This tool loads the Puter.js library, which routes AI calls through Puter's user-pays infrastructure — no API keys required on your side and no account setup on this site.

Images are processed as base64 data URLs sent directly from your browser to the selected AI provider via Puter. They are not stored on this website's servers.

Everything else — folder reading, metadata editing, CSV/JSON export, EXIF embedding, ZIP packaging — runs 100% in your browser.