Photo Assisted Keywords: 100% Local AI Auto-Tagging for RAW Photos on Mac

Hi Product Hunt! 👋 I'm Maik, the solo developer behind Photo Assisted Keywords. For years, coming home from a long photo walk with my Sony Alpha 7R II meant one thing: hours of agonizing, manual data entry. Staring at hundreds of massive RAW files and typing "architecture," "summer," or "pedestrian" into Capture One was killing my creative energy. I wanted an AI to do this, but I completely refused to upload my massive, uncompressed RAWs to any cloud service or pay a monthly subscription for it. So, as a passionate "tiny old Xcoder" closing in on 63, I decided to build the solution myself. Photo Assisted Keywords leverages Apple's MLX framework to run a massive 12-Billion parameter AI model (Gemma) entirely locally on Apple Silicon. Here is what it does: Local RAW Processing: It analyzes heavy formats (ARW, CR3, NEF, DNG, etc.) directly on your Mac. No cloud, 100% privacy. Smart Context: It grabs EXIF data (time/date) to give the AI context (e.g., knowing it's "sunset" in "summer"). Lightning Fast "Fuzzy" Logic: It detects similar images in burst shots and instantly copies keywords to save massive amounts of compute time. Pro Workflow Ready: It writes standard XMP sidecar files (Dublin Core & Lightroom schemas) that instantly sync with Capture One, Lightroom, or Photo Mechanic. Building this pushed me to my limits, and getting a personal congratulatory email from Tim Cook for this very app back in May was a surreal moment I’ll never forget. I built this to get my creative time back, and I hope it helps you do the same. If you want a deeper look at the tech behind the local MLX integration, check out the YouTube link in the gallery! I’ll be here all day to answer your questions and talk about Mac development, AI, and photography! 📸