Organize your photos effortlessly with AI-powered intelligence

XMP (Extensible Metadata Platform) is a standard developed by Adobe for embedding or attaching metadata to image files. A sidecar file is a companion file stored alongside your original image — for example, a file named "IMG_0042.RAF" would have a companion named "IMG_0042.XMP" sitting in the same folder. Because many raw image formats (like Fujifilm RAF, Phase One IIQ, or Hasselblad 3FR) are proprietary and not meant to be modified directly, applications like Lightroom® and Capture One® write your ratings, labels, keywords, and descriptions into these XMP sidecar files rather than touching your originals. This keeps your raw files pristine and allows multiple applications to share the same metadata. Talos: Photo Assistant reads and writes to these same XMP sidecar files, meaning that AI-generated tags and descriptions flow seamlessly into your existing Lightroom® or Capture One® workflow without duplicating or moving your photos.

Yes — Talos: Photo Assistant can work with photos stored on a NAS, but there are a few things to keep in mind for the best experience. Connection speed matters most. A 10 Gigabit wired Ethernet connection is ideal and will provide a seamless experience comparable to local storage. If your NAS or Mac supports USB-C direct-attach storage, that is also an excellent option. Standard Gigabit Ethernet is workable for moderate libraries but may feel sluggish when Talos is scanning or caching metadata across a large number of files. Drive type is also important. NAS units with spinning hard drives (HDD) can be noticeably slower than SSD-based storage due to higher seek times and lower random-read throughput. If possible, choose a NAS that includes an SSD cache tier, or store your active photo library on an SSD-based volume. This makes a significant difference when Talos is reading thumbnails and metadata for many images at once. Wi-Fi is not recommended for large libraries. Wireless connections introduce latency and bandwidth variability that can make Talos feel slow. A wired connection is always preferable.

Talos: Photo Assistant uses a secure cloud connection to deliver its AI features — including smart tagging, image descriptions, and the photography coach. The good news is that the bandwidth requirements are modest. Any internet connection with a symmetrical speed of 10 Mbps or faster (both upload and download) will be sufficient for smooth operation. Most modern broadband, cable, fiber, and even many LTE connections comfortably exceed this threshold. Talos sends only resized copies of your images to the API for processing — not your full-resolution originals — so even on a modest connection the uploads are quick. Your original photos never leave your Mac.

Talos: Photo Assistant requires macOS 14 (Sonoma) or later and is optimized for Apple Silicon. Any Mac with an M1, M2, M3, or M4 chip — including the MacBook Air, MacBook Pro, Mac mini, Mac Studio, and Mac Pro — is more than powerful enough to run Talos smoothly. Apple Silicon Macs provide excellent performance for the on-device tasks Talos handles locally, such as reading and writing XMP metadata, caching thumbnails, and managing your library index. The AI processing itself happens in the cloud, so raw GPU or Neural Engine performance has less impact than you might expect — a base M1 MacBook Air performs just as well as a high-end Mac Pro for Talos's workload. Intel-based Macs are not supported at this time.