Photo Body Analysis
Upload a full-body photo and we'll estimate your bone proportions — shoulder-to-hip balance, overall elongation, and torso-to-leg ratio — to suggest which Kibbe families fit you best.
The pose-detection model downloads once and runs entirely in your browser. Your photo never leaves this device, is never transmitted, and is deleted the moment you close this page.
Upload a full-body photo
Drag & drop it here, or choose a photo from your device. Everything is processed locally — nothing is uploaded.
For best results
- Stand straight, arms relaxed at your sides, facing the camera.
- Wear fitted clothing so your shoulders, hips, and legs are visible.
- Capture your full body — from head to ankles.
- Use good, even lighting and a plain background.
- The photo is analyzed entirely on your device and never uploaded.
Privacy first: the pose-detection model runs inside your browser, so your image never leaves this device. This is a starting-point estimate of bone proportions — not a full Kibbe typing.
How to Use Photo Analysis
Getting a reliable result is mostly about the photo itself. Follow these five steps.
Choose your photo
Pick a clear, front-facing, full-body photo from head to ankles. Stand straight with your arms relaxed at your sides — the algorithm needs to see your shoulders, hips, and legs.
Upload or capture
Drag & drop a photo, choose one from your device, or capture one with your camera. JPG, PNG, and WebP are all supported.
Wait for analysis
The first run downloads the pose-detection model (about 8.5 MB, cached in memory for the session). After that, analysis completes in a few seconds.
Review your results
You get your best-matching Kibbe family, a percentage-ranked top 3, and your measured proportions — shoulder-to-hip, elongation, and torso-to-leg.
Go deeper
Open the full styling guide for your type, combine your result with the text quiz, or rotate the 3D landmark view to inspect the detection.
Pro tip
The most reliable result comes from combining methods: take the text quiz as well. If you already have, this tool blends your answers with the photo automatically.
Understanding Your Results
Four numbers explain how the tool arrived at your match.
Shoulder-to-hip
1.00 means perfectly even. Above 1, your shoulders read wider than your hips (a yang-leaning clue); below 1, your hips read wider.
Overall elongation
Your height divided by your shoulder width. Higher values point to a longer, more vertical frame — the signal behind the “vertical line” concept.
Torso-to-leg
Compares torso length with leg length. Lower values mean proportionally longer legs, which contributes to a lengthened silhouette.
Pose confidence
How reliably BlazePose found all key landmarks. Aim for 90%+. Poor lighting, partial crops, or baggy clothes lower the score.
How Photo Body Analysis Works
This tool uses Google's BlazePose pose-detection model (via TensorFlow.js) to locate 33 landmarks on your body inside your browser. From those points it measures the relationships that matter most in the Kibbe system: your shoulder-to-hip ratio, your overall height-to-width elongation, and your torso-to-leg ratio. Those measurements are then mapped against all 13 Kibbe types to find the best matches.
What It Can and Cannot Tell You
A photo can reveal skeletal proportions well, but the full Kibbe system also weighs facial features, the softness of your flesh, waist definition, and absolute height — none of which a single photo can judge. Treat this as a strong starting point, then confirm with the text quiz and the detailed type guides. For the most accurate result, do both — the two methods combined are noticeably more reliable than either alone.
The 3D Landmark View
After analysis you can open an interactive 3D Landmark View. This is a simple visualization of the 33 pose points that BlazePose detected — not a body scan or 3D model of you. It lets you rotate the skeleton to see how the detection algorithm saw your pose. If your browser cannot run 3D, the 2D skeleton overlay is used instead.
Your Privacy
The pose model is downloaded once to your device and cached in memory for the session. Photos are processed with WebGL on your own hardware. No image data, measurements, or results are sent to any server, logged, or stored — see our Privacy Policy for details.
Photo Analysis FAQ
Answers to the questions people ask most often.
Is my photo uploaded to a server?
No. The entire analysis runs inside your browser using WebGL on your own hardware. Your photo is never transmitted to a server, never logged, and never stored — it is cleared from memory when you close the page. The only network request is downloading the pose-detection model itself.
Do I need to create an account?
No. Photo analysis is free and requires no sign-up. Everything happens locally on your device.
What kind of photo gives the best results?
Use a front-facing, full-body photo from head to ankles with your body fully in frame. Stand straight with arms relaxed at your sides. Fitted clothing makes your shoulders, hips, and legs much easier to detect — loose or baggy clothes reduce accuracy. Good, even lighting and a plain background also help.
Why does it download a model the first time?
The pose detector (Google BlazePose) is about 8.5 MB. It downloads once on your first analysis and is cached in memory for the rest of the session, so subsequent analyses are instant. Nothing about your photo is sent during this download.
How accurate is photo analysis?
Photos are excellent for measuring skeletal proportions — the relationships between shoulders, hips, and torso. But the full Kibbe system also weighs facial features, flesh softness, waist definition, and absolute height, none of which a single photo judges well. Treat this as a strong starting point, then combine it with the text quiz for the most reliable result.
Why did analysis fail on my photo?
Common causes: your full body was not in the frame, you were too far or too small in the image, the lighting was very poor, your browser has WebGL disabled, or you were in an unusual pose (e.g., arms crossed or a side angle). Retake the photo following the how-to steps above and try again.
Which devices and browsers are supported?
Any modern browser with WebGL — Chrome, Edge, Firefox, and Safari, on both desktop and mobile. You can capture a photo directly with your phone camera. Very old or restricted browsers (including some private or in-app webviews) may not support WebGL.
What does the 3D Landmark View show?
It visualizes the 33 pose points BlazePose detected, rendered as a rotatable 3D skeleton. It is a visualization of the detection — not a body scan or a 3D model of you. If your browser cannot run WebGL, the 2D skeleton overlay is used instead.
How is photo analysis different from the text quiz?
Photo analysis measures your bone proportions objectively from an image. The quiz relies on your own assessment of your frame, flesh, and facial features. The two methods complement each other — combining them (55% photo, 45% quiz) is noticeably more reliable than either alone, and the tool does this automatically if you have taken the quiz.
Can I use a photo of someone else?
The tool is designed for self-analysis. Running a photo of someone else would report their proportions, not yours, and it would not take their consent into account. For the most useful and respectful result, use your own photos only.
What next?
Your photo analysis is a strong first read. Complete the picture by taking the text quiz, then explore the full styling guide for your best-matching type.