Search by Concept, Not by Name
Type "business presentation with charts" or "sunset at the beach" — PhotoCore finds the right image even if the file is named DSC_0312.jpg. Finder cannot.
Finder is the backbone of macOS file navigation — excellent for moving, copying, and renaming files. But it is completely blind to image content. If you can't remember the exact filename, that photo is gone for good. PhotoCore integrates with your existing folders and adds local AI-powered semantic search — without changing how you use Finder at all.
A complete overview of the differences between macOS Finder and PhotoCore for managing and finding images.
| Feature | macOS Finder | PhotoCore |
|---|---|---|
| AI semantic search | Keyword-based only 1 | Yes — natural language queries |
| Understands visual content | Superficial (OCR + macro-tags) 2 | Yes — via SigLIP 2 / CoreML |
| Visual similarity search | No | Yes — instant vector search |
| AI smart rename | Sequential rules only | Yes — from image content |
| Offline external disk archives | Only when connected | Yes — search while disconnected |
| NAS / network volumes (SMB, NFS) | Spotlight blind to visual content on NAS | Full — local DB, paths on NAS |
| Indexing control & progress | Automatic, unpredictable (Spotlight) | Manual trigger, real-time progress bar |
| Search by filename | Yes | Yes (and by visual content) |
| Data privacy | High (local) | Maximum — 100% local + CoreML |
| Cost | Free (included in macOS) | Free (Light) / €29.99 one-time |
1 Finds simple objects (e.g. "cat", "car") via Apple Visual Lookup, but does not understand complex phrases, contextual descriptions or multi-element compositions.
2 Recognises text in images via Live Text (OCR) and applies standard macro-category tags (people, places, animals). It does not compute semantic vector embeddings.
PhotoCore doesn't replace Finder — it supercharges it with a layer of visual intelligence that macOS doesn't have on its own.
Type "business presentation with charts" or "sunset at the beach" — PhotoCore finds the right image even if the file is named DSC_0312.jpg. Finder cannot.
Finder renames with sequential numbers (photo_001, photo_002…). PhotoCore analyses each image and suggests a descriptive name: sales-flow-diagram.png instead of IMG_7843.png.
Found the right shot? One click and PhotoCore analyses visual vectors to show all aesthetically and conceptually similar images. Finder has no equivalent feature.
Index an external disk once. PhotoCore saves thumbnails and vectors locally: you can search and browse even when the disk is in a drawer.
Finder reads the name, date, and EXIF metadata. It has no idea whether an image shows a sunset, a business diagram, or a presentation cover. PhotoCore computes a 768-dimensional visual embedding for every photo via SigLIP 2 on the Apple Neural Engine — it sees what you see, regardless of the filename.
Search across 10,000 photos in <50ms — instant results via Accelerate.vDSP.
PhotoCore imports nothing, creates no separate library, and moves none of your files. It simply points to your existing folders — on internal disk, external drive, or NAS — and indexes everything locally. Finder keeps working exactly as before, now with the power of semantic search added on top.
Perfect compatibility: PhotoCore reads the same folders you manage in Finder. Rename a file in Finder? PhotoCore updates it on the next scan. Move a folder? Just re-index. No conflicts, no library to keep in sync.
Spotlight indexes filenames and EXIF data on SMB/AFP volumes — but never computes visual embeddings on network paths. That means it can't find a photo by what it shows on a NAS, only by what it's called. PhotoCore solves this at the architecture level.
PhotoCore's SQLite database (photocore.db) always lives on your Mac, inside ~/Library/Application Support/PhotoCore/. Only the file paths point to the NAS. If the network drops for a moment, your local index is completely safe — no corruption, no data loss. Contrast this with Spotlight, which never builds visual representations for remote volumes at all.
Spotlight on your Synology: finds project-photo.jpg by name — search "blue wireframe diagram" returns nothing.
PhotoCore on the same NAS: "blue wireframe diagram" instantly surfaces the right image across your entire library — all AI inference done locally on your Mac's Neural Engine.
Spotlight indexes on Apple's schedule — only when your Mac is idle and plugged in. Many users report that files added two days ago still don't appear in search. PhotoCore puts you in charge.
Trigger indexing any time. A real-time progress bar shows exactly how many images have been processed.
PhotoCore runs a concurrent IndexingActor in Swift: indexing is non-blocking, the UI stays responsive, and you always know the exact state. Trigger it on demand — perfect after importing a batch of new photos or connecting a fresh external drive.
Photographers and professionals often store years of work across multiple external drives. With Finder and Spotlight, those archives vanish the moment a drive is unplugged. PhotoCore changes that completely.
Connect each drive and let PhotoCore index it. Thumbnails (256 px HEIC) and visual vectors are saved locally on your Mac. The drives can go back in the drawer.
On the train with no drives attached, search "construction site at sunset" — PhotoCore surfaces the image with a preview, even though the disk is at your office.
Every result shows the origin volume and full path. You know the file is on "Archive 2022 — WD 4TB" before you open a single drawer.
Finder and Spotlight forget a drive the instant it's disconnected. No preview, no path, no result. Your archive simply disappears from macOS until you reconnect.
Download PhotoCore for free. The Light version is free forever up to 200 images. Full access unlocks with a one-time purchase — no subscription ever.