Private photo tool, no account required

Free AI Color Analysis for Your 12-Season Palette

Choose one natural-light selfie to measure a 12-season match, build a personal palette, and see the limits behind the result.

  • Free MVP analysis
  • No sign-up
  • Temporary private processing
  1. Upload
  2. Analyze
  3. Result
Drop a selfie here, or choose a photoJPG, PNG, or WebP up to 15 MBFor a more reliable result
  • Indirect natural daylight
  • No filter or heavy makeup
  • Cheeks and forehead visible
  • Neutral top and background

A better photo makes a better result

Four checks before you upload

Filters and artificial light change color temperature. For your AI color analysis, use a recent, unedited photo that represents how you look day to day.

Face natural daylight

Stand near a window and avoid direct sun or mixed indoor lighting.

Keep your face visible

Look straight at the camera with hair away from your cheeks and forehead.

Skip filters and makeup

Use the original photo with light or no makeup and no beauty filters.

Choose a neutral setting

A plain background and neutral top reduce reflected color around your face.

See the guide before you decide

A report built for the next time you get dressed

The free result explains the match. The saved guide turns it into practical choices for shopping, outfits, makeup, hair, and the colors you already love.

  • A free season result before checkout
  • A complete palette organized for shopping
  • Outfit, beauty, hair, and jewelry decisions
  • A phone-friendly guide you can reopen later
Natural-light portrait of a brunette woman beside a sample Warm Autumn color analysis
Natural-light portraitPaid guide sample

Season match

Warm Autumn

Undertone

Warm leaning

Guide focus

Everyday decisions

Start your shopping shortlist here

Wear this tomorrow
Start with camel and olive, then bring muted teal closer to the face for a calm, low-contrast outfit.
Keep a color you already love
Black does not have to leave the wardrobe. Open the neckline and add warm metal or an earthy color near the face.
Beauty and hair direction
The full guide narrows lipstick, blush, jewelry, and hair-color choices without pretending to measure features the current engine cannot see.

Save the finished guide to your phone and open it in the fitting room, at the makeup counter, or before a salon appointment.

This licensed Unsplash portrait is used for interface preview. The adjacent Warm Autumn report is illustrative; it is not a diagnosis or evidence of accuracy.

Your result comes before checkout

See your result free. Pay only when you want to use it beyond today.

Start with your season, color direction, representative palette, and one useful recommendation. Then decide whether to keep the complete shopping, outfit, beauty, and hair guide.

Free color result

$0

Complete the analysis without an account and see enough detail to decide whether the result is useful to you.

  • Primary and alternate season match
  • Representative personal palette
  • Metal guidance and a short explanation
  • Photo quality and result context

Full Color Report

One-time purchase, no subscription

$9.99

for one report

Turn one analysis into a saved decision guide for getting dressed, shopping, choosing makeup, and discussing a hair-color change.

Everything in your full guide

  • A 36-color shopping palette organized by how you will use it
  • Outfit combinations you can use without starting from scratch
  • Clear clothing, makeup, hair-color, and jewelry decisions
  • Why your primary direction leads the alternate result
  • A phone-friendly reference for the fitting room or makeup counter
  • Permanent account access with print and PDF export
Get my free result first

One-time checkout appears after you receive an eligible color analysis result.

Inside the MVP

What this AI color analysis actually measures

The tool separates photo screening from color scoring. Your browser checks that exactly one face is centered and front-facing, then resizes and converts the photo on your device. After you consent and start, only the full normalized pixel image is uploaded privately for sampling on the forehead, cheeks, and jaw.

Accepted pixels are converted from sRGB to CIELAB, a color space designed to compare perceptual color differences. A versioned scoring map then places the measured lightness, warmth, and chroma near one of 12 seasonal profiles. The top-two gap is a score separation, not a probability or accuracy percentage.

One-face check on your device

MediaPipe checks face count, size, centering, head roll, and basic frontal geometry before the upload button can run the analysis.

Four measured skin regions

The current engine samples the forehead, both cheeks, and jaw. It does not infer a result from demographic labels, hair color, or eye color.

CIELAB color relationships

Median color, region spread, luminance spread, and Delta E consistency checks reduce the effect of isolated pixels and flag unreliable photos.

Versioned 12-season scoring

The same measurements produce the same score under the same engine version. The MVP reports the engine and calibration versions with every result.

Current limitation: season-engine 0.1.0 uses an uncalibrated prototype map. It is useful for repeatable product testing, but it has not yet been validated against a diverse expert-draping benchmark. The page therefore makes no accuracy guarantee.

How it works

From selfie to palette in three steps

The first version keeps the workflow intentionally short. You can review the result, retake the photo, or delete the analysis.

01

Upload one clear selfie

Choose a photo in natural light and confirm temporary face-photo processing.

02

Measure color relationships

The engine compares normalized cheek and forehead color signals across the 12-season model.

03

Review your palette

Get your season match, alternate, palette, metals, explanation, and photo-quality notes.

Choose the right approach

Photo tool, self-guided quiz, or in-person draping?

Each method answers a slightly different need. A fast online result is most useful when its trade-offs are visible instead of hidden behind an accuracy claim.

ApproachBest forMain trade-off
This photo-based MVPA private, repeatable starting palette with no account or appointment.Camera white balance, exposure, nearby colors, and makeup can shift the measurements.
Self-guided questionnaireExploring color ideas without uploading a face photo.Answers depend on how accurately you judge your own undertone, contrast, and features.
In-person fabric drapingComparing real fabrics under controlled light with an experienced consultant.Usually takes more time, costs more, and quality varies by consultant and lighting setup.

Test the result

How to check whether your palette is stable

Treat one AI color analysis result as a hypothesis. Repeating the photo under controlled conditions is more informative than trusting a confident-looking label from one image. Retake whenever the lighting setup changes.

Why two photos may disagree

  • Automatic white balance and phone processing can warm, cool, brighten, or smooth skin tones.
  • Direct sun, mixed bulbs, strong shadows, and tinted window light change the sampled color relationships.
  • A bright wall, colored top, hair, or phone case can reflect color onto the cheeks and jaw.
  • Beauty filters, portrait mode, makeup, and exported screenshots can alter texture or color even when the change looks subtle.

A practical four-step check

  1. 01Start with a recent, unedited image beside a window, wearing a neutral top against a plain background.
  2. 02Run a second photo from the same session, then repeat on another day in similar indirect daylight.
  3. 03Look for recurring palette direction. If the top two seasons remain close, use both as candidates rather than forcing certainty.
  4. 04Before an expensive wardrobe or hair-color decision, compare physical swatches or book an in-person draping session.

Before you upload

Useful answers, without the sales pitch

AI color analysis is a styling aid, not a medical or identity assessment. Lighting and camera processing can change the result.