VisionTagger overview

VisionTagger helps you make photo libraries searchable and consistent by generating rich, structured metadata from images entirely on your Mac. It runs on Apple Silicon (M1 or later) and macOS 26 (Tahoe) or later, using local vision models — so your images or metadata don’t need to leave your machine. You can download preconfigured vision models in-app or bring your own by linking a GGUF vision model + matching GGUF projector. Built-in metadata sections include Title, Description, Keywords, Content & Style, and Safety & Compliance, and you can extend these with custom sections/fields. Each custom field supports a data type (Boolean, Text, or List of Texts) and its own prompt, so you can define a metadata schema that matches your workflow and conventions.

For output, VisionTagger can:

Create XMP sidecars and/or embed metadata via ExifTool (installed separately).
Export JSON or TXT per image or as a single file for a batch.
Write metadata back to your Photos Library.
Apply Finder tags.

It supports common image formats (including JPEG, PNG, TIFF, HEIC, WebP) and can process images from folders or directly from your Photos Library. A free trial lets you process up to 100 images (no time limit), and the app is sold as a one-time purchase (this major version).