The whole shoot, handled.
Weddings, events, portraits, pets — four stages, one platform. Here's what runs under the hood, shown with real screenshots.
AI selects your keepers
Every frame is graded 0–100 across seven dimensions — Clarity, Focus, Open Eyes, Warmth of Emotion, Composition, Energetic Moment and Exposure — then combined into an overall score and a Pick / Alternate / Review recommendation.
- Duplicate & burst grouping collapses near-identical frames into one decision.
- Person-aware: different people are never merged as duplicates.
- Runs on-device by default — your shoot never has to leave your machine.
The blink, fixed — from the burst itself
When the best frame of a burst has someone mid-blink, LightVision borrows that person's open eyes from another frame in the same burst — aligned, harmonized and identity-verified. The donor is your client's own frame, never a generated face.
- Works on group shots — each blinker is fixed from their own best frame.
- Identity gate: it never borrows the wrong person's eyes.
- Non-destructive — the borrowed eyes are saved as a layer on the keeper you can undo at any time.
★ Best of 4 Eyes open Eyes closed · 22% 👁 Eyes fixed · Best take Detecting a blink on the keeper…
Edited in your style, automatically
LightVision reads the develop settings from your finished work and gives every new import a matched first-pass edit by scene and lighting. It learns your full grade — tone curves and an 8-band color mixer, not just exposure sliders — and keeps a separate profile per shoot type, so your wedding look never bleeds into corporate work.
- Learns from your XMP sidecars, DNG settings or delivered JPEGs.
- Tunes to the actual lighting — what you do to high-ISO, backlit or warm-white-balance shots.
- You stay in control — it's a head start, not a hijack.
The face rescuer for dim rooms
Receptions and ceremonies are shot at high ISO, and faces come back soft and grainy. LightVision's on-device face restore is built to rescue exactly those faces — it measures each face's sharpness and noise, only recommends restoration where degradation is real, then restores texture and detail.
- Live in the editor with a 0–100 strength slider and per-face masking.
- Restores detail and texture rather than lifting exposure — brightening stays a separate develop step.
- Non-destructive, so you control how much of the restored face blends back in.
Face restored High-ISO original Find anyone, anything, any moment
Once a shoot is in, LightVision understands it — by face, by meaning and by moment — all on-device, no cloud, no API key.
Pick a face, find every shot
Powered by on-device face recognition, LightVision groups everyone by who they actually are — robust across pose, lighting and expression. Right-click any face and choose "Show all photos of this person" to pull every shot they appear in across the shoot.
- Right-click any face to name the person or tag a role — Bride, Groom, Best man — and the name sticks to their identity as the shoot grows.
- Runs fully locally — your guests' faces never leave your Mac.
- Works on RAW files too, reading the camera's embedded preview.
Every shot One face Everyone covered — down to the family group shot
Set "Require 1, 2, 3 or 5 in Picks of each person" and LightVision flags anyone who falls short. Then define must-have groups — "Couple + bride's parents" — and it checks that a frame with the whole group together actually landed in your Picks, before the client ever asks.
- Per-person progress — 2 of 3 in Picks · needs 1 more — with the biggest gaps sorted first.
- One click — Add best — fills each gap with the top-ranked frames you haven't picked yet.
- Face matching runs on-device; guests' faces are never uploaded.
3 of 3 · covered 2 of 3 · short Guests find their own photos — privately
Turn on "Let guests find themselves by face" — strictly opt-in and off by default — and a delivered gallery gains a "Find my photos" picker. A guest taps their own face and the grid filters to just the shots they're in, on any device.
- Matching runs entirely in the guest's browser — nothing is sent to any server during a search.
- Only anonymous face vectors are published — no names, no identities — and nothing loads until the gallery password is unlocked.
- One honest limit: it's a deterrent, not encryption, and it can occasionally miss a face.
Just theirs All photos Search in plain English
Type "kids laughing" or "golden hour by the lake" and LightVision finds the shots that look like that — not just ones tagged that way. It's the default search, ready the moment photos are imported. One-tap preset chips — kissing moments, candid laughs, group photos — build themed collections instantly, and any search saves as a live collection.
- Right-click any photo and "Find visually similar" to pull every lookalike.
- No grading, no pre-tagging, no waiting — it works on import.
The shoot, sorted into moments
LightVision automatically sorts each shoot into named moment collections — Getting Ready, Ceremony, First Kiss, First Dance, Cake Cutting, Speeches, Dancing — tuned to the shoot type, then builds a chronological Story Arc that walks your best picks through the day in order.
- Every shoot gets its own moment folders plus a per-scene breakdown, rebuilt automatically as photos are graded.
- Tuned to the job — wedding, event, corporate, portrait or product each get their own named moments.
Type a word, get a mask
Add an AI Select mask, type a region in plain words — "the sky", "her dress", "the dog" — and LightVision returns a pixel-accurate selection, no brushing. It's open-vocabulary, so you can name almost anything in the photo, not a fixed list of skies and subjects.
- Runs entirely on your Mac — no cloud, no uploads.
- The text-made mask becomes a normal editable adjustment layer — refine it with a click, then dial exposure, color and more on just that region.
- Click to add or Option-click to subtract. First use needs a one-time model download.
Masked & edited Type "the dog" Client galleries that wow
Publish a polished client gallery with favorites and email-gated downloads — on your own domain, in your brand.
Turn galleries into print sales
Switch on the built-in print store. Clients order wall art, albums and prints straight from their gallery — in your brand's colors — and a pro lab prints and ships every order. Manage it all from the photographer portal.
Sell the work — don’t give it away
Turn on protection and your images stay yours until a client pays. Previews carry a watermark baked into the pixels — not an overlay a right-click can peel off — while the full-resolution files stay off the page entirely, released one at a time to the client who bought them. Included-download counts are enforced on the server, so a private window doesn't reset them.
One honest caveat: no gallery online can stop an OS screenshot — which is exactly why the watermark is baked into whatever a screenshot captures.
Feature questions, answered
How does LightVision cull thousands of photos?+
It grades every frame 0–100 on seven dimensions — clarity, focus, open eyes, warmth of emotion, composition, energetic moment and exposure — then sorts them into Pick, Alternate and Review, with duplicate and burst grouping so near-identical frames become a single decision.
Can it edit photos in my own style?+
Yes. LightVision reads your real develop settings from finished work, builds a per-shoot-type profile, and applies a matched first-pass edit to new imports by scene and lighting. You stay in control of the final look.
Can it fix a blink in the best shot?+
Yes — Best Take borrows that person’s open eyes from another frame in the same burst, blends them onto the keeper, and verifies it is the same person so it never uses the wrong eyes. It works per person in group shots, the donor is always your client’s own frame — never a generated face — and it’s a layer you can undo anytime.
Can it rescue faces shot in poor light?+
Yes. On-device face restore is built for the soft, grainy faces that dim receptions and high-ISO ceremonies produce — it measures each face’s sharpness and noise, only suggests restoration where degradation is real, and gives you a 0–100 strength slider per face. Non-destructive, so you control the blend.
Can I find a specific guest across the whole shoot?+
Yes. Right-click any face and choose “Show all photos of this person” to pull every shot they appear in. Recognition runs on-device — no cloud upload of your guests’ faces — and holds up across pose, lighting and expression.
Can I make sure I deliver enough shots of each person?+
Yes. Set “Require 1, 2, 3 or 5 in Picks of each person” and the People hub flags anyone who falls short, biggest gaps first. One click adds each missing person’s best unpicked frames until the gap closes — all matched on-device, so guests’ faces never leave your Mac.
Can I name people and tag wedding roles?+
Yes. Right-click a face to type a name or tag a role — Bride, Groom, Best man and more. Names bind to the face’s identity rather than a throwaway “Person 7,” so they stick as the shoot grows, and they save with your project.
Can I check I got the family and group shots?+
Yes. Define must-have groups from people you’ve named — “Couple + bride’s parents” — and LightVision checks whether a frame with the whole group together made your Picks. “Add best” fills any gap with the highest-scoring together-frames. Matching is best-effort face recognition, so give the results a glance.
Can wedding guests find their own photos in the gallery?+
Yes, and it’s off by default. Flip on “Let guests find themselves by face” per gallery and guests tap their own face to filter to their photos — matching runs entirely in the guest’s browser, and only anonymous face vectors are published, never names or identities. Nothing works until the gallery password is unlocked.
Can I search my photos in plain English?+
Yes. Meaning-based search runs on-device — type “kids laughing” or “golden hour by the lake” and it ranks your library by what the pixels actually show, not by tags. It’s the default search and works the moment photos are imported. Any search saves as a live collection.
Does it organize the shoot into moments?+
Yes. LightVision automatically sorts each shoot into named moment collections — Getting Ready, Ceremony, First Kiss, First Dance, Speeches, Dancing — tuned to the shoot type, and builds a chronological Story Arc that walks your best picks through the day in order.
Can I make a selection mask by typing?+
Yes. Add an AI Select mask, type a region in plain words — “the sky”, “her dress”, “sunglasses” — and LightVision returns a pixel-accurate selection, no brushing. It’s open-vocabulary, runs entirely on your Mac, and becomes a normal editable adjustment layer.
Can I deliver galleries and sell prints from the same app?+
Yes. Publish polished client galleries with favorites and email-gated downloads, and switch on the built-in print store — a pro lab prints and ships every order.
Spend your weekends shooting, not culling.
LightVision is in beta — join free, no card required, and we'll help you get set up.