








faces-h learns every face in your photo library — on your device, without a server, subscription, or account.
Pick any folder on your computer. faces-h scans it in the background — no GPU needed, no waiting around before you can use it.
The on-device InsightFace model clusters appearances automatically. You name the people you recognize — the app handles the rest.
Every photo of Sarah, every shot with David and Emma together — one click. Your whole library organized by the people who matter.
Click a face cluster in the sidebar, open Name this person, and type a name — every photo of them is labeled instantly. Typing a name that already exists offers to merge the two clusters instead, so nobody ends up duplicated. One-photo clusters wait under Single-face clusters until you name or merge them.
When the app isn't confident enough to assign a face automatically, it lands in the To review queue (the badge in the sidebar). Each card shows the best guess — favoring someone you've already named over an anonymous cluster when scores are close, since that's the suggestion you can actually act on. Confirm it, pick someone else, mark it Not relevant, or skip. Confirming shows a brief "Looking for more matches…" indicator while the app re-checks the rest of your library for the same person — often resolving several more faces at once. The app never silently guesses: anything below its confidence threshold waits for you.
In a photo's detail panel, hover a wrongly-labeled face and click This person is wrong. The correction re-evaluates the whole cluster in the background — faces that no longer fit are pulled out for review automatically.
Pick one or more people to find every photo they appear in together, optionally within a date range. Switch to exact people mode to find photos of only those people — no one else in the frame.
Add more folders any time (Ctrl+O) — network drives and NAS shares included — and rescan the library with Ctrl+R. Manage Scan Folders… (··· menu) lists every folder you've added with its status, and lets you stop scanning one — photos already found there are never removed, only future scans of that folder stop. Photos deleted from disk hide themselves on the next scan and come back automatically if the file returns. Browsing and scanning never modify your files — the only file action the app can take is the explicit “Move to Recycle Bin” delete you confirm yourself (e.g. from Find Blurry Photos), always recoverable from the Bin.
Find Blurry Photos uses a live cutoff slider — drag it and watch the grid refilter so you see exactly which shots each level catches. Find Duplicates groups exact copies and same-shot re-saves, each copy listed with its folder, filename, and size, with a one-click “keep one per group”. Before anything is deleted, a confirmation shows every file’s thumbnail, name, folder, and size plus the total space freed. Every file — local or network — is backed up inside the app first, then goes to the Windows Recycle Bin when possible; folders without a Recycle Bin fall back to a safe, already-backed-up removal, so it feels the same everywhere.
Rotate Sideways Photos finds shots that are upright only in tag-aware viewers, plus photos where no face was detected because the image is actually sideways — a quick background scan checks those against every angle. Each suggestion shows a before/after preview so you see the exact result first. Applying it rewrites the original file in place; the pre-rotation version is always backed up inside the app first, the same way for local and network files, so nothing is ever silently altered for good.
Export Named People saves a small file with your named identities — no photos inside. Import it on another machine or library and matching clusters get their names back automatically.
faces-h checks for new versions automatically and stays quiet unless one is found — or check anytime from Help > Check for Updates…. Installing runs quietly in the background and restarts the app for you; it skips the old "uninstall the previous version?" prompt entirely.
The activity log at the bottom shows what the engine is doing live (drag its verbosity
slider for more or less detail). Full logs live in
%APPDATA%\com.faces-h.app\logs. Power users can tune recognition
thresholds, detection filters, and log verbosity in
%APPDATA%\com.faces-h.app\config.json — first launch after installing
takes a minute or two while the face engine warms up.
Free and open source. Always.
Windows installer is code-signed. Certificate provided by SignPath Foundation.