Free tool

Which of your photos should go first?

Add up to 6 photos of the same person. Each one is scored on the things a photographer would look at, then ranked, so you can see which one earns the first slot and what is holding the others back.

Before you add them

  • Same person in every photo. This compares shots, not people.
  • The face needs to be visible. Sunglasses and heavy shadow give it nothing to read.
  • Mix the angles. Six near-identical photos produce six near-identical scores.
  • Any ordinary photo file. JPG, PNG, WebP, and iPhone HEIC, which is converted on your device.

0 of 6 photos added

Who is in the photos?

This only switches off the wrong penalties, like marking down a shaved head or expecting makeup.

What are the photos going for?

How to pick the photo that goes first

On a dating profile, a company page or anywhere else a face appears next to a name, the first photo does almost all of the work. People decide whether to keep looking in about the time it takes to scroll past. The difficult part is that you cannot see your own photos the way a stranger does. You know which one was taken on a good day, which one your friend liked, and which one you have simply looked at the most. None of that is what a stranger sees in a second and a half.

That is the gap this tool is for. It scores each photo you give it against the same fixed list of photographic checks, then ranks them. The ranking is useful, but the reasons are more useful: they tell you what the weaker shots were losing points for, which is the part you can actually fix next time you pick up a camera.

What it is actually judging

Every check here is about the photograph, not the person in it. That distinction is the whole design. A photograph can be badly lit, badly framed, shot from too low or taken half a second after a smile has faded, and all four of those are fixable. None of them say anything about the face.

So the tool never scores how someone looks. It will not reward a stronger jaw, softer features or direct eye contact, because those are preferences dressed up as measurements. It scores the execution: the angle the camera was held at, how the light fell, whether the background is competing with the face, whether the expression reads as intended.

The checks, and why each one is on the list

Camera angle covers how the lens sits relative to the face. Front-on, three-quarter and profile can all score well. What is penalised is distortion: a phone held too close enlarges whatever is nearest it, which is usually the nose, and that is a property of the lens rather than the face.

Height and distance is the most common quiet failure in phone photos. Held below chin height, a phone looks up the face and flattens the jaw. Held at eye level and stepped back half a metre, the same face in the same light reads completely differently.

Chin and jaw scores position, never size or shape. A chin brought slightly forward and then slightly down separates the jaw from the neck and lets the jaw edge catch light. A chin pulled back compresses the neck, a chin raised high shows the underside of the face. Both are a two-second adjustment.

Where the eyes go and Expression are scored together in spirit. A level gaze into the lens creates immediate connection. Looking away works when it looks like there was a reason to look away, and reads as a missed moment when there was not. On expression, the thing that matters is whether the eyes are present. A small smile with engaged eyes beats a wide grin with flat ones, every time, and a neutral face is only weak when it looks empty.

Light and Shadows are where most ordinary photos lose their points, and also where they are easiest to recover. Soft light from in front, a window on an overcast day, open shade outdoors: all of these are free and all of them beat an overhead bulb.

Framing and crop, Background and Sharpness are the photograph as an object. Is anything cut off distractingly, is there a lamp growing out of the head, is the focus actually on the eyes. These are the checks people skip and then wonder why a good moment produced a weak picture.

Grooming and styling is assessed only as photographed, and only for how it works with the light and the angle. No hairstyle is better than another. A shaved head, a full beard, makeup or no makeup can all score full marks.

Balance is off by default, and that is deliberate. Reading symmetry needs a front-facing, eye-level frame with a neutral expression and no head tilt, which is to say a passport photo. Run against ordinary portraits it mostly asks you to take a different photo, which is not an answer anyone wants.

How to read your result

Start with the gap, not the order. Scores within a few points of each other are the same score. The tool tells you when the top photos are tied, and when it does, the honest advice is to pick whichever one you prefer: there is no measurable difference between them and your own instinct about which one looks like you is worth more than a point and a half.

Then read the weakest sections of your best photo. That is the most valuable line in the whole result, because it describes a photo that is already working and tells you the one thing that would make it work harder. A shot scoring 84 with its lowest marks in background and framing is a shot that needs the same pose taken two steps to the left.

Finally, look at what the bottom photos have in common. If four of your six lose points on light, the problem is not those four photos, it is where you take photos. That pattern is the real output of running 6 shots through at once rather than one at a time.

What the scores are and are not

The scores are subjective judgements made by an AI model against a written rubric, not measurements. Run the same set twice and individual scores move by a few points. That is why ties are reported as ties, and why no claim is made that a 78 is really better than a 76. What is stable across runs is the shape of the result: the clearly strong photos stay at the top, the clearly weak ones stay at the bottom, and the reasons stay consistent.

It is also worth saying what no tool can do. It cannot tell you whether a photo looks like you on a normal day, whether it suits the person you are hoping will see it, or whether the story it tells about your life is the one you want told. Those are the questions worth thinking about once the technical ones are out of the way.

Common questions

How many photos should I add?

Between four and 6 is the useful range. Two works and tells you very little. The value comes from seeing a pattern across several, and from having enough variety that the pattern means something, so vary the angle, the light and the setting rather than adding six frames from the same burst.

Does it keep my photos?

No. The photos are combined into one image in your own browser, sent once for scoring, and never written to our storage. If you are not signed in, the result is held in your browser only and clears itself after 24 hours. Signed in, your results stay on your account.

Why does it ask whether the photos are of a man or a woman?

Only to switch off the wrong penalties. The two versions differ in what they refuse to mark down, such as a shaved head or an absence of makeup, and in two genuinely different photographic details, hair-cast shadows and beard presentation. Nothing else is scored differently, and the choice is yours rather than something guessed from the photo.

Can I use it for a work headshot rather than a dating profile?

Yes. Every check on the list is a general portrait check. Set the goal to strong and dramatic or natural and candid rather than warm and approachable, and read the result the same way.

Where to go next

If one side of your face keeps coming out better, the Better Side tool measures that from a single frontal photo. If the problem is the pose rather than the picking, Pose Maxing works on a photo you already have. For the whole profile, bio included, there is the dating profile review.

For the background reading: how to pose for photos covers the adjustments this tool keeps suggesting, and picturemaxxing explains why the photo you like least is often the one that performs best.

Reconnecting…