Can you really find your twin somewhere in the world?

In everyday use, a doppelganger is an unrelated person who looks strikingly like you: the stranger who could pass for your sibling in a group photo. People phrase the search as finding their twin in the world or as finding people who look like them, but it is one task: compare a clear photo of your own face with as many other faces as possible and look at the closest results.

Nobody can promise you a twin. The saying that everyone has a set number of doubles is folklore, not a measured fact. What is well documented is that unrelated lookalikes exist: Canadian artist François Brunelle began photographing pairs of them in 1999, and scientists later used his portraits to study why the resemblance happens.

Whether you find a convincing double depends on chance and on how many photos of you both a tool can see.

How does software decide that two faces look alike?

Most face-similarity tools find the face in a photo, line it up so the eyes, nose and mouth sit in standard positions, and convert it into a long list of numbers, often called a face embedding or faceprint. Two faces are compared by measuring how close their numbers are.

The result is a similarity score, and each tool decides for itself what counts as close. A 92% in one app cannot be compared with 92% in another, and it is not the chance that two photos show the same person.

Researchers separate two tasks. One-to-one matching checks whether two photos show the same person, as when a phone unlocks for its owner. One-to-many searching compares one face against a large collection and returns the closest candidates; a lookalike search is the second kind, so it hands you the most similar faces whether or not any of them is you.

The US National Institute of Standards and Technology (NIST) calls it a false positive when software wrongly treats photos of two different people as the same person. In a lookalike hunt, those near misses are the whole point. Anywhere identity matters, they are the danger.

Why do unrelated people look so alike?

Part of the answer is arithmetic. Faces vary along a limited set of features, such as face shape, eye spacing, nose and mouth, and among billions of people some combinations repeat closely enough to fool a friend or an algorithm.

Biology plays a part too. In a 2022 study in Cell Reports, researchers took 32 lookalike pairs from Brunelle's photographs and measured their likeness with three face recognition algorithms. All three grouped sixteen pairs together, and nine of those sixteen shared a large set of common genetic variants while differing in DNA methylation and microbiome.

The authors noted clear limits: a small sample, two-dimensional black-and-white images and mostly European participants. The pairs were unrelated, so sharing some variants is not the same as being family.

Photos exaggerate resemblance as well. Similar angles, lighting, hairstyles, glasses, filters and low resolution can make two faces look closer in a picture than side by side in person.

Where can you find people who look like you?

Start with your own face and with services that explain clearly what happens to it. A clear selfie can become a faceprint that recognizes you in other photos: in 2021 the US Federal Trade Commission alleged that the developer of the Ever photo app switched face recognition on by default for most users and used faces from users' photos to build its technology. The settlement required express consent and the deletion of face embeddings made without it and of any models built with users' photos.

So before you upload anywhere, look for plain answers: who can see your photo, whether your face joins a searchable database, how long it is kept, whether it trains the system and whether any payment renews. Then choose a route:

  • Art matching: the Art Selfie feature in the Google Arts & Culture app compares your selfie with faces in artworks from partner museums. In January 2018 Google said it did not use the selfie for anything else and kept it only while searching for matches; check the app's current notice before taking part.
  • Lookalike communities: Twin Strangers, for example, lets you register and upload a straight-on photo of your face for free. Its FAQ explains that registering means other users can find you and that automated matching runs on paid credits.
  • Reverse image search on your own photo: general engines react to the whole image, so expect matching poses, outfits and backgrounds more often than true face twins.
  • Public-web face search on yourself: a face search such as FaceSonar compares your face with faces on public web pages, so people who resemble you can appear alongside any photos of you.
  • Photo projects and lookalike contests: join only when the organizer explains how your photo will be used and published.

Use a photo of only your own face, with no other people, no children and nothing in the background that shows where you live.

What does a FaceSonar lookalike search show, and what doesn't it show?

FaceSonar is a public-web face search rather than a lookalike game, but running it on your own face is a direct way to see who resembles you online. You upload one photo, which is stored privately on FaceSonar's server and never made public, choose the face, and a short on-screen scan prepares it. That scan is only a preparation step and never shows or claims a match.

Unlocking the report runs the real search against a third-party face index of public web pages. The report lists pages where a similar face appears, each with a similarity score from 0 to 100 and a link to the source, so you can see every photo in context. Reports are paid and unlocked with a search pack or a plan, bought by card, and photos behind unpaid searches are deleted automatically after 30 days.

  • It shows public web pages where a face similar to yours appears, each with a link you can open and check.
  • It scores every result: a higher score means the faces are closer by the index's measure, not that the person is you or anyone in particular.
  • It can include your own photos if they are online, next to people who only resemble you.
  • It does not cover private accounts, messages, dating-app profiles behind a login or live video streams, so a short list does not mean you have no lookalike.
  • It does not tell you who anyone is: whatever a page says about the person in it is that page's claim, not a confirmed fact.
  • For a lookalike search, upload only your own face; FaceSonar is not for searching for or identifying children.

Every lookalike in a report is a real person who did not ask to be found by you: do not message or follow them, and never share their photo without their consent.

Why is a lookalike match never proof of identity?

A lookalike search is built to return the closest faces it can find, so it usually produces candidates even when none of them is you. That is exactly why resemblance should never decide who someone is.

NIST's testing shows the risk. A false positive in a one-to-many search puts the wrong person on a list of candidates for further scrutiny, and NIST found that most algorithms it tested had different error rates for different demographic groups, with large variation between algorithms.

In everyday life the harm is smaller but real: a stranger blamed for a post because they resemble its author, or a dating match dismissed as fake because they look like a model. Before acting on a resemblance, look for independent context, such as the same photo rather than a similar face, and never confront or accuse anyone on a face alone.

Is it your lookalike, or is someone using your photo?

Sometimes the twin turns out to be you. If a result shows your actual photo, with the same background, pose and clothing, it is a copy rather than a lookalike, and on a profile under another name it may be impersonation.

A reverse image search can help settle it: identical or edited copies point to reuse, while a different photo of a similar-looking person points to a genuine lookalike. If your photo is used without permission, save screenshots with the page address and date, report the profile to the platform as impersonation, go to the police if it is used for fraud or threats, and do not engage with the account. Our guide to fake profiles made with your photos walks through each step.

How do you share a lookalike find without causing harm?

Finding someone who looks like you is fun; turning them into content without asking is not. Posting their picture next to yours can bring them comments, tags and attention they never wanted.

  • Share a lookalike's photo only with their consent, even if it is already public; asking is realistic only where people chose to be found, such as fellow members of a lookalike community.
  • If you cannot ask, keep the find private, or share only your own photo and describe the resemblance.
  • Do not tag them, follow them, message them or try to work out who they are and where they live.
  • Never use a lookalike's photo to pose as them, or your resemblance to pose as someone else.
  • Leave children out entirely, including comparisons involving other people's kids.
  • If someone asks you to take a post down, take it down.

A resemblance is a coincidence you share with a stranger, not permission to involve them.

Questions people ask

Does everyone have a doppelganger?

No one can say for sure. The idea that everyone has a set number of doubles is a saying, not a finding. Unrelated lookalikes clearly exist, but whether your closest match is a near twin or a passing resemblance is largely luck.

Are doppelgangers related to each other?

Not in the family sense. A 2022 Cell Reports study found that some extremely similar, unrelated lookalike pairs shared many genetic variants, which helps explain the resemblance. The study was small, though, and sharing some variants does not make two people relatives.

What does a similarity score or percentage mean?

It comes from that tool's own model and shows how close two faces are by its measure. Different tools use different models and thresholds, so the numbers are not comparable, and a high score is not the probability that two photos show the same person.

How can I find people who look like me online?

Search with a clear photo of your own face. Art-matching features compare you with museum portraits, lookalike communities match you with members who chose to be found, and a public-web face search such as FaceSonar lists public pages with a similar face, each scored and linked in a paid report.

Can FaceSonar find my twin anywhere in the world?

No tool can compare your face with everyone on Earth. FaceSonar searches a third-party face index of public web pages, so it cannot see private accounts, messages, dating apps behind a login or live video. Lookalikes can appear in a report, but every result is a lead to review, never proof of who someone is.

Is it okay to contact my lookalike or post their photo?

Only with their consent. A stranger whose face appears on a public page did not agree to be found, and sharing their picture next to yours can bring them unwanted attention, tags and messages. If you cannot ask, keep the find to yourself or share only your side.

Sources and further help

These links go to official or independent resources. FaceSonar is not affiliated with them.