How to choose culling software: five questions

Most culling advice tells you which tool is best without asking what you shoot. This does the opposite. Five questions, in order, each one eliminating part of the field, then a decision path and a one-week test that settles the last two. No ranking, and no pretending the answer is the same for everyone.

Stop asking which one is best

Every thread about culling software ends the same way. Someone asks which tool is best, twenty photographers name the one they bought, and nobody says what they shoot. The tools in this category are not better and worse versions of the same thing. They disagree about who should make the decision, where the work should happen, and what you should be paying for.

So here is a different approach. Five questions, in order. Each one eliminates part of the field, and by the fifth you are usually choosing between two products instead of seven. Answer them honestly, including the ones where the honest answer is not flattering.

This page is the method. If you want the raw specifications, every tool laid out on platform, price, detections and handoff, that lives in the culling software spec table, and the longer prose walkthrough of each product is in our guide to the best photo culling software.

Question 1: when do you actually do this work?

Not when you intend to. When it actually happens. For most working photographers the answer is Sunday night, or Tuesday between shoots, or the evening after a full day when you already know you should be asleep.

It identifies which bottleneck you are buying against, and there are two very different ones.

If the software is the slow part, if you are already at the desk and the pain is Lightroom building previews while you wait, then every AI tool here fixes that and you can move on to question two. This is a genuinely solvable problem and the tools are good at it.

If getting to the desk is the slow part, be honest that no purchase in this category changes that today. Aftershoot, Narrative Select, FilterPixel, Photo Mechanic and Lightroom Classic are all desktop programs. Imagen runs in a browser but wants your files uploaded first. Buying any of them still means the job begins when you are home with the cards ingested. It is worth knowing which problem you have before you spend $120 a year solving the other one.

A UK survey of more than 300 wedding photographers, reported by PetaPixel, put culling at about 11 percent of working time and editing at 55 percent. It predates these tools, so read it as a shape rather than a measurement. If your evenings are being eaten by editing, the best culling tool in the world moves a small number.

Question 2: do you want a decision, or faster evidence?

This is the real fork in the category, and it is a question about your temperament rather than your genre.

Some tools hand you a finished selection to audit. Aftershoot works this way: it proposes the set, you check its work, and you are reviewing a decision that has already been made. When you trust it, this is enormously faster. When you do not, auditing a stranger's choices feels worse than making your own.

Some tools refuse to decide and make the evidence instant instead. Narrative Select works this way. Focus scores on every frame, a panel of magnified faces so you can see the one guest blinking in a group of twelve, and near identical frames grouped together with the sharpest surfaced. You still make every call, just far faster than you could unaided.

Neither is correct. Photographers who like being the author of every choice are miserable auditing an AI's list, and photographers who want the first pass off their plate find the evidence approach exhausting. You will know which one you are within an hour of a trial, and both companies give you one free.

If your honest answer is "I am not sure I trust software with my clients' faces at all", that is a legitimate position and worth thinking through properly before you buy anything. We went through it in whether you can trust AI culling.

Question 3: can your files leave your machine?

One tool in the mainstream field is cloud based. Imagen runs in a web interface with nothing installed, which means your photos are uploaded before any work happens. Aftershoot, Narrative Select and FilterPixel all process locally on your own computer.

Three things decide this one for you, and none of them are about quality.

Your internet. A wedding is commonly 2,000 to 4,000 RAW frames. On ordinary home broadband, that is an upload you plan your evening around, not a step you barely notice. If you shoot on location for days at a time, or your last hotel wifi experience still hurts, local processing wins before the feature comparison starts.

Your contracts. Corporate, government, school and some editorial work carries restrictions on where image files can be sent and stored. If you have ever signed a clause about third party processors, read it before you upload a job.

Your operating system. Every tool named here runs on both macOS and Windows, so this rarely eliminates anyone, but performance and GPU behavior are not identical across the two. We compared the practical differences in culling software on Mac versus Windows.

Question 4: does the pricing model match your year?

Photography income is lumpy. Pricing models are not, and the mismatch is where people end up paying for months they did not shoot.

Prices checked August 2026 from each vendor's own pricing page.

Two traps worth naming. The first is the headline tier. Narrative's $10 Lite plan does not include Scenes View or the Close-Ups panel, the two features most reviews praise, so a culling buyer is really comparing $240 a year against Aftershoot's $120. The second is the add-on. Imagen's $12 culling price is real, but it attaches to an editing subscription, so it is only a $12 decision if you were buying the editing anyway.

Then put the number in proportion. The Knot's 2025 study of over 10,000 US couples put the average wedding photographer at $3,000. A $120 a year subscription is four percent of one booking, so price is rarely the thing that should decide this.

Question 5: what has to happen after the picks?

The last question is the one people skip, and it undoes more purchases than any other. A tool that picks brilliantly and hands off badly costs you the time it just saved.

If you edit in Capture One, the field shrinks immediately. Aftershoot and Narrative Select both hand off to it. Imagen's site talks about Adobe Creative Cloud catalogs and does not mention Capture One at all, and FilterPixel lists Lightroom Classic support. That single fact settles the choice for a lot of people before they read a word about accuracy.

If you live in a Lightroom catalog, ask how the picks arrive. Star ratings, flags and color labels behave differently in a catalog, and for RAW files the marks usually go into XMP sidecar files rather than changing the original. Aftershoot's documentation is explicit that it does not overwrite originals. Get this right before you commit, because rebuilding a selection by hand because the labels landed in the wrong place is a genuinely bad evening. The mechanics are in getting selections into Lightroom cleanly.

If a second shooter, an assistant or a studio is involved, ask who else needs to see the picks. Narrative includes 2 user seats on Premium and 4 on Ultra. Photo Mechanic's license covers one user on up to two computers.

The decision path, in one table

Answer the five questions, find your row, then run the trial. This is a starting point, not a verdict.

Prices checked August 2026 from each vendor's own pricing page.

How to settle it in one week

Shortlist two, never five. Then run this, because a trial spent poking at features tells you nothing.

Use a shoot you already culled by hand, ideally six months old, one where you still remember which frames you chose. You need a known answer to measure against. A fresh shoot is useless for this because you have no baseline.

Look only at the disagreements. Frames the software picked that you did not, and frames you picked that it passed over. Everything else is noise. Ten minutes in that pile tells you whether the tool's taste is compatible with yours, and there is a fuller protocol in how to test AI culling accuracy on your own work.

Time the whole loop, not the analysis. Card to shortlist to editor, including the export. Vendors advertise the fast part. You are buying the whole loop.

Check the tier your favorite feature lives on before the trial ends. Narrative's free trial is of Ultra, the top plan, which is generous and slightly dangerous: you can spend two weeks loving features that are not on the plan you intend to buy.

Full disclosure, since it belongs here rather than buried: we are building a culling tool, native for iPhone, iPad and Mac, still being built, free for founding members while we build. Our booking page is live and free today. If you need a working answer this Saturday, pick a row from the table above and run the trial. That is the same advice we would give a friend.

FAQ