Read the energy
The photo is downscaled in memory and the analyzer measures local contrast to find where the eye-catching detail sits.
Drop in a photo. The analyzer measures where the visual energy sits and scores how well the rule of thirds, golden ratio, diagonal method, and a centred grid match the frame — entirely in your browser. Nothing is uploaded.
The Composition Analyzer reads a photo on your own device, builds a map of where its contrast sits, and scores how closely that energy aligns with four framing systems — the rule of thirds, the golden ratio, the diagonal method, and a centred grid. It draws the best fit over your image and links to it as a live overlay. It is a transparent heuristic, not a verdict.
No machine learning and no upload. Just contrast, geometry, and a fit score you can reason about.
The photo is downscaled in memory and the analyzer measures local contrast to find where the eye-catching detail sits.
It checks how much of that energy falls near each system's lines and points — thirds, phi, diagonals, and centre — and turns it into a fit score.
The top grid is drawn over your image. Open it as a live overlay to refine the crop on your own reference.
A fast read on the structure already in your frame — useful for choosing a crop, confirming an instinct, or teaching composition.
Thirds, golden ratio, diagonal, and centred — ranked by how well each aligns with your photo.
The winning grid is rendered over your image so you can see where the lines land.
The nearest root rectangle, so you can also consider the dynamic-symmetry overlays.
The photo is read locally with the browser's file reader. There is no upload and no server.
It measures contrast against geometry — no opaque model deciding your photo is good or bad.
Open the matching overlay live on your reference to crop and recompose, free and local.
A photographer, a teacher, and an editor on using the fit scores as a second opinion.
It is a sanity check before I commit a crop. If thirds and golden both score low, that usually tells me my subject is dead-centre — sometimes on purpose, sometimes not.
Students argue about composition in the abstract. Dropping their shots in and seeing the grid land makes the conversation concrete in seconds.
For a photo edit I run a batch through it to spot which frames already sit on a strong structure. It never decides for me, but it narrows the pile.
The analyzer is deliberately simple and honest about its limits. Here is exactly what it measures and how to read it.
Every digital image is a grid of brightness values. Where neighbouring values differ sharply — an edge, a silhouette, a bright subject against a dark ground — the eye is drawn. The analyzer approximates that by measuring local contrast across a downscaled copy of your photo, producing a map of visual energy. It then asks a single geometric question of each composition system: how much of that energy sits close to this system's lines and points? The rule of thirds is scored against its two-thirds lines and four intersections; the golden ratio against the 0.382 and 0.618 divisions; the diagonal method against the corner diagonals; and the centred grid against the middle. The system whose structure best coincides with the busy parts of your frame scores highest. The rule of thirds and golden ratio glossary entries cover the systems themselves.
What the score is not is a measure of quality. A photograph can ignore every grid and still be superb — symmetry, dead-centre framing, and large fields of negative space are all deliberate choices that score low against thirds. A high score means your contrast happens to align with that system; a low score means it does not. Read it as a description, never a grade. The negative space guide and the broader composition overlays for photographers pillar explain when breaking the grid is the right call.
The tool can tell you where the contrast in your picture falls. It cannot tell you whether that was a good idea. That part is still yours. — Grid Maker Pro studio notes
Because the analysis is a transparent heuristic rather than a trained model, it behaves predictably: the same photo always returns the same scores, and you can reason about why. That is a deliberate trade. A neural network might "feel" more authoritative, but it would also be a black box, would likely need a server, and would break the promise that your image never leaves your device. We chose the version you can understand and that stays entirely on your machine.
Four kinds of frame and the grid the analyzer tends to favour for each.
Anyone choosing a crop or studying a frame. Each links to a workflow guide.
It downscales your photo, measures local contrast to build a map of where the visual energy sits, and then scores how much of that energy falls near each system's lines and points — the thirds lines, the golden divisions, the corner diagonals, and the centre. The grid that best coincides with the busy parts of your frame scores highest. It is a transparent heuristic, not machine learning.
No. The analyzer reads your image with the browser's file reader and processes it on an in-memory canvas. No bytes are ever sent to a server — the same privacy guarantee as the rest of Grid Maker Pro. Your reference image stays on your device.
No, and it is not meant to be. Composition is a judgment, not a measurement. The analyzer surfaces which grids align with where your photo's contrast already sits — useful for choosing a crop or confirming an instinct — but a deliberately centred or broken composition can be excellent while scoring low. Treat the scores as a prompt, not a verdict.
Four common framing systems: the rule of thirds, the golden ratio, the diagonal method, and a centred composition. It also reports your aspect ratio and the nearest root rectangle so you can consider the dynamic-symmetry overlays too.
Because the analyzer only measures where contrast falls relative to each grid's lines. Many excellent images use negative space, symmetry, or a centred subject that does not sit on a thirds line. A low score means the photo does not follow that particular system, not that it is poorly composed. See the negative space guide.
Yes. Paintings, illustrations, posters, and screenshots all work, because the analyzer only needs contrast to measure. It is a quick way to check whether a layout or artwork sits on a recognised compositional structure.
It reports whether your frame is closest to a root-2, golden, root-3, or root-5 rectangle. Those proportions have their own dynamic-symmetry overlays — reciprocal diagonals and armatures tuned to the rectangle — which can suit the frame better than a generic thirds grid.
No. After the page loads, the analysis is local and instant. Because nothing is uploaded, it works the same on a slow connection or offline once the page is cached.
Free forever. No signup. The photo is read and analysed entirely on your device — nothing is uploaded.
Analyze a photoOther free utilities in Grid Maker Pro: