What is FaceOracle?

An entertainment service that reads the mood, style, and impression of a single photo

What we're actually making

FaceOracle is a small AI report that puts your photo's impression into words. It's not a golden-ratio scoreboard or a beauty grading service — the goal is to turn the vibe of a photo into useful personal-color, hair, and makeup ideas.

AI Technology

FaceOracle uses Anthropic Claude AI's multimodal vision. The model interprets the image directly to generate results based on the face's mood and visual features. It is not used for face recognition (faceprints) or identity verification. The uploaded photo is sent to the external Claude API for analysis and then discarded rather than stored on our servers; only the shareable result is kept temporarily for 24 hours. See How it works and the Privacy Policy for the full data flow.

What it's actually good for

• Picking a new profile photo and seeing which shot reads better at a glance
• Pulling color, hair, and makeup ideas from a single selfie
• Checking the visual chemistry between a couple of photos for fun
• Starting the day with a one-liner from the daily fortune

What it isn't for

• Decisions where someone's life is on the line — hiring, admissions, lending — it isn't a basis for that.
• Identity verification, access control, or biometric authentication. Not the right tool. Cannot be used for those.
• Medical diagnosis or precise age and health calls — it doesn't do that.
• Rating, comparing, or making fun of someone else. Please keep it light.

Who we are

FaceOracle is run by a one-person team in Korea. All product, design, code, and editorial decisions are handled in-house, and emails to our contact address are answered by the team directly. We do not invent expert personas or reviewers.

How the content is made

Every blog post is planned, written, and reviewed by me personally. Here are the standards I follow, and the things I deliberately do not do.

✓ What I do

• Pick genuinely useful, practical topics — photography, styling, personal color — grounded in things I have shot, worn, and compared myself.
• Stamp every post with a publish date and a last-reviewed date, and bump the review date whenever I revise it.
• Cross-link related posts so a topic can be read in more depth, not as a dead end.
• When something is found to be wrong, fix the body of the post and note it, rather than quietly deleting it.

✕ What I refuse to do

• Invent statistics, studies, or expert names. I write from general knowledge and first principles.
• Claim that a face determines personality, ability, health, or identity.
• Rank looks as better or worse, or write anything that puts a person down.
• Publish raw AI drafts — I re-check the facts, wording, and safety of every piece before it goes live.

More to read

• → How it works — the analysis pipeline
• → Editorial policy — how we write
• → Content principles — what we say and don't
• → Operator's note — why I run it this way
• → Content hubs
• → The Face-Reading Literacy Room — culture & psychology
• → FAQ

Contact

For questions, deletion requests, or feedback, email yuseong2099@gmail.com.