How to Check If Someone Is Real Online: 9 Safer Steps

To check if someone is real online, first review what they are asking you to do, then verify only the claims that matter through independent public sources. Exact-image search, candidate public pages, and an optional in-app video call can add context, but none proves identity, account ownership, intentions, or safety. A no-result search is inconclusive.

1. Slow the interaction down. Urgency makes weak evidence feel stronger than it is.
2. Check for harmful requests first. Money, cryptocurrency, gift cards, passwords, one-time codes, identity documents, secrecy, and unfamiliar links are reasons to pause.
3. Preserve only the necessary context. Keep the profile URL, username, relevant claims, dates, and messages without building a broad personal file.
4. Choose one claim that matters. Decide what you actually need to know before collecting more information.
5. Run exact-image search. Look for copies of the same photo and inspect the original page context.
6. Add public-web face search only if necessary. Use it when candidate pages across different photos could add useful context to a specific decision.
7. Open every promising source. Compare publisher, caption, date, username, and context instead of relying on a thumbnail or score.
8. Offer an optional in-app video call. Use it only as a limited consistency check when both people are comfortable.
9. Stop at resolution or uncertainty. Choose a reversible next step, and do not keep searching to force a verdict.

If the person is already asking for money, credentials, remote access, or secrecy, further image research is unnecessary. Refuse the request and use official block, report, financial-provider, or support channels. You may pause or end the interaction without proving deception.

If a limited photo question remains, start with the guide to finding public source pages from a photo, then compare the roles of face-search tools.

Before paying, inspect the real saved 100-candidate sample report and the published source-review methodology. For first-meeting decisions, use the wider dating-safety hub.

Keep only the context needed for a safety decision: the profile URL, username, claims that affect your choice, relevant messages, dates, and any payment or threat information. Do not build a broad personal file. Leave private images, identity documents, addresses, relatives, and unrelated people out of the review, and use platform tools where possible.

Urgency can outrun good judgment. If the person is already pushing for secrecy, emotional intensity, or money, that matters as much as anything you may find in search.

Separate claims into three groups: checkable through an independent source, private but not needed, and unsafe to request. You may check a public job or event through its official website. You do not need a home address, private family detail, or identity document for an early dating decision. Record only conflicts that affect the immediate choice. Memory errors or privacy limits are not proof, and a consistent story still does not guarantee good intent.

Before opening a candidate, decide which source could answer the claim. An official employer page may support a public job claim; a repost or people-search directory may add little. Note the date and whether sources are independent. Do not collect contact details, relatives, addresses, or private-account clues merely because they appear nearby.

Start with exact-image search to look for copies of one photo. Add public-web face search only when pages with different photos could clarify a claim you need to check. Open every original source. Neither search reaches private accounts, identifies who controls a profile, guarantees coverage, or turns an empty result into proof that an account is real.

1. Start with one clear photo.
2. Add face search only if necessary. Keep any returned page unconfirmed and treat no result as inconclusive.
3. Save the source pages that look relevant.

221B covers the public-web part of this workflow. It may surface public pages, profiles, articles, and websites with a similar-looking face, but you still need to inspect those pages and leave account ownership unresolved.

According to Google Search Help, image-search results may include similar images and pages containing the same or a similar image. Exact-image search can trace a photo, but it cannot show who copied it or who runs a profile. Face search may surface other public pages, yet lookalikes, old pages, reposts, or stolen photos can explain a resemblance. Describe the outcome as reuse, resemblance, a candidate source, or no useful result.

Use the face search and image search comparison to choose the least invasive method. Stop after reviewing enough original pages to answer the narrow question or leave it unresolved. Repeating the same search across services can expose the photo again and duplicate weak evidence without adding useful context.

Review each candidate page for its publisher, date, caption, image use, and the public claim it supports. A similar-looking face, different name, or conflicting timeline is something to check, not proof of fraud. Do not contact candidates, employers, relatives, or unrelated account holders. Leave weak evidence unresolved.

• Name context: note how a public page uses a name without assuming that resemblance proves identity.
• Timeline: compare city, work, school, and activity dates only when they affect the claim you need to check.
• Photo pattern: image variety or polish cannot tell you whether an account is genuine or deceptive.
• Source context: public captions and dates may clarify a claim, but do not map relatives, contacts, or hidden accounts.

Prefer official or first-hand public sources over scraped directories and reposts. A company biography, established publication, platform help page, or named event page usually gives more context than an anonymous profile directory. Several pages copied from one source are not independent support. If two sources conflict, keep both explanations open and ask whether the conflict changes your next safe action.

For a relevant candidate, write one plain sentence about what the page shows and one about what it cannot show. Check only the claim that affects your decision. Do not guess someone's relationship status, character, criminal conduct, or intent from public photos. If the context is unclear, keep it unclear and choose a safer action that does not require naming the person.

Worked example: what does a bounded review look like?

The saved 221B public-figure case preserves the first 100 usable candidate pages from 299 parsed candidates across 27 canonical hostnames. The goal is not to confirm a name. It shows how a reviewer opens a source, records what the page supports, leaves control of the account unresolved, and stops before resemblance becomes identity proof.

1. Open one candidate page. Check that it still exists and that the relevant image appears there.
2. Record only what the source supports. A page may support a date, event, caption, or public role; it does not show who controls another account.
3. Keep alternatives open. Reposts, copied thumbnails, old pages, and lookalikes can explain a result.
4. Stop at the boundary. If the source cannot answer the narrow question, leave it unresolved.

This is one saved query, not a performance benchmark or a claim of complete coverage. Opening the sample runs no live search, uploads no photo, and uses 0 credits.

An optional in-app call may add limited context when both people agree. It cannot prove a legal name, background, intent, or future conduct. Someone may decline for valid privacy or safety reasons. Never demand a new proof image, a forced pose, an identity document, an intimate image, or anything that reveals a home or current location.

• Live interaction: movement and conversation can add current context, although lag, filters, recorded media, and AI-generated video can still mislead.
• General appearance: you may notice whether someone broadly resembles the profile without demanding a pose or treating resemblance as identification.
• Boundary behavior: the call may show whether the other person respects consent and ordinary limits, but that is not a character or safety verdict.

Treat the call as one optional check that both people agree to. It cannot show who controls an account, prove a legal name, confirm a job or relationship status, reveal someone's background, or establish intent. Refusing a call is not proof of deception. If doubt remains, choose a reversible next step instead of asking for more proof.

Use the dating platform’s own call and verification features when available instead of moving to an unfamiliar service. The FBI IC3 warning on dating-verification schemes describes links that collect personal or payment information. Open official safety features yourself and never enter card details through a link supplied by a match. A platform badge means only what that platform says it checks; it is not a background check or personal-safety guarantee.

Start with conduct that can cause harm: requests for money, cryptocurrency, gift cards, passwords, one-time codes, identity documents, remote access, secrecy, unfamiliar verification links, or rapid emotional pressure. These signals justify pausing, blocking, or reporting without deciding who the person is. A polished profile or search result cannot cancel out unsafe behavior.

• Money or device access: payment, investment, gift-card, cryptocurrency, bank-transfer, or remote-access requests.
• Account access: passwords, one-time codes, card details, identity documents, or unfamiliar verification links.
• Pressure and control: secrecy, urgency, threats, repeated contact, sexual pressure, or anger when you set a boundary.

Australian eSafety guidance cites a 2022 study in which 72.3% of surveyed dating-app users reported online sexual harassment, aggression, or violence from someone they met through a dating platform in the previous five years. These survey results describe reported experiences; they do not predict whether any one person is safe.

Reported experience in that study	Share of surveyed users
Repeated contact after saying no	47.3%
Unwanted sexually explicit message	47.2%
Pressure to meet in person	34.5%
Threats	18.9%

A call or profile-photo comparison remains a limited check. If the profile image may be synthetic, use the guide to checking an AI-generated dating profile photo without treating a visual clue or empty search result as proof.

A request for money, cryptocurrency, gift cards, passwords, codes, or remote access already gives you a safer next step: stop. The FBI's 2025 IC3 report recorded US$929,287,469 in reported confidence and romance losses. Scamwatch's combined 2025 data listed A$139.9 million in romance-scam losses in Australia.

These are reported loss totals, not the odds that a specific account is fraudulent. They show why a money request matters more to your next action than a face-search result.

The FTC's 2025 report data says nearly 60% of people who reported losing money to a romance scam said it began on social media. The FTC's romance-scam guidance advises people not to send money or gifts to an online love interest they have not met in person.

1. Do not send more money or information.
2. If you paid, contact the bank, card issuer, exchange, or payment service through its official channel.
3. Preserve the relevant messages and use official platform and government reporting tools.

When the evidence is unclear, leave it unresolved and choose a step you can reverse. Stop sharing information or money, use the platform's block and report tools, and tell someone you trust. If money, threats, or intimate images are involved, save the relevant evidence and contact the payment service, a support service, or an official reporting channel promptly.

• Do not send money.
• Do not agree to secrecy because they ask.
• Do not explain away repeated conflicts because the conversation feels good.

If needed, talk it through with someone outside the situation. Distance from the conversation can make a pattern easier to see.

For a first meeting, follow current local guidance such as the Australian eSafety Commissioner’s online-dating advice: choose a busy public place, arrange your own transport, tell a trusted person, keep control of your phone and departure, and avoid sharing your home address early. In immediate danger, contact local emergency services. A search is only one optional input; your boundaries, platform tools, financial protections, and a practical meeting plan matter more than a tidy set of public pages.