Photo Library

Photo Library adds AI-written titles, descriptions, and keywords to a folder of photos, without changing the images. It is built into the desktop app and does not require Python, a source checkout, or a Lightroom plugin.

Index a folder

1. Open More → Photo Library and choose a folder.
2. Connect OpenAI or OpenRouter in the same view.
3. Review the scan: image count, total size, and any camera formats that need an extra decoder.
4. Run a small test first. Include a mix of shots (for an event: day and night, wide and close, blurry, signage, and at least one unrelated image).
5. Index the full folder.

You can stop at any time and reopen the job later from Recent archives. Photos already done with the same provider and model aren’t charged again.

What leaves your machine

Your files stay where they are. For each photo, HivemindOS sends one display-sized preview to the provider you chose, never the whole folder. That provider’s own terms cover processing and retention.

Review results

Uncertain results go to a review queue. You can edit titles, descriptions, keywords, and Lightroom keyword hierarchy before approving. The model is told not to guess who people are or their sensitive traits.

Export to Lightroom or another photo manager

• Export safe copy saves approved XMP sidecar files inside HivemindOS. Use this to test importing.
• Write beside originals asks you to confirm, then writes a sidecar next to each photo. It only updates sidecars HivemindOS created and never replaces your own XMP files. Metadata is never written into the image itself.

Test a small export in Lightroom, Peakto, or your photo manager before writing to a large archive.

Supported formats

JPEG, PNG, WebP, and GIF work everywhere. macOS handles more formats with its built-in decoder. On Windows and Linux, install ImageMagick from the decoder setup prompt for RAW, HEIC, TIFF, or BMP files.