Is This a Scam?
Paste a suspicious text, email, or link. Get a plain answer.
This tool provides AI analysis, not definitive security verdicts. When in doubt, contact the organization directly.
Free. No signup. We don't store what you paste.
Your message is analyzed and discarded. Nothing is saved, logged, or used for training. This tool runs on Ask Safely's privacy architecture: the same auto-delete, no-training, encrypted infrastructure that powers the full product.
How it works
1. Paste the message exactly as you received it.
2. The checker analyzes it against known scam patterns, URL structures, impersonation signals, and persuasion tactics.
3. You get a verdict, an explanation of what triggered it, and what to do next.
It catches the patterns that are hard to spot now that AI-generated scams no longer have typos.
What this catches
Package delivery scams. "USPS: Your package could not be delivered due to an incomplete address." Real USPS tracking texts only come from the short code 28777, only after you opt in, and never include a link. Any USPS text from a regular phone number with a link is fake. Every time.
Bank fraud alert scams. "CHASE FRAUD ALERT: Did you authorize a Zelle payment of $487.00?" Your bank never asks you to reply YES or NO by text. They never ask for one-time passcodes. They never tell you to transfer money "to yourself" for safekeeping.
IRS and government impersonation. "IRS Notice: Your 2025 tax refund of $3,847.00 has been approved." The IRS never initiates contact by text, email, or social media. First contact is always a mailed paper letter. If it arrived on your phone, it's not the IRS.
Tech support pop-ups. "Your computer has been compromised. Call this number immediately." Real error messages from Apple, Microsoft, or Google never include a phone number. The number in the pop-up goes to a scam call center.
Subscription renewal scams. "Your Netflix/Amazon Prime/antivirus subscription has expired. Click here to renew." These emails mimic real brands but the links go to lookalike domains (netflix-billing-update.com instead of netflix.com).
AI-generated phishing. The new generation of scam messages has no typos, no awkward grammar, and can be personalized with your name and details scraped from public sources. The old advice ("look for spelling errors") no longer works. The checker analyzes persuasion patterns, URL structure, and impersonation signals that survive even perfect English.
The numbers
The FBI logged 22,364 AI-related fraud complaints for the first time, with $893 million in losses. Voice cloning alone accounted for an estimated $352 million in elder fraud.
AI-generated phishing emails are now 24% more effective than human-written ones. In 2023, they were 31% less effective. AI closed that gap in two years.
The old heuristics — bad grammar, weird formatting, "this doesn't sound right" — are being defeated by design. That's why this tool exists.
What real messages look like
USPS only sends tracking texts from the short code 28777, only after you text them a tracking number first. The texts never contain a link and USPS never charges a redelivery fee.
The IRS contacts you first by paper mail, sent to your address on file. They never demand immediate payment by phone, text, or email. They never threaten arrest.
Your bank sends fraud alerts through the banking app or by calling the number on the back of your card. They never ask you to reply YES/NO by text, never ask for your password or one-time code by message, and never tell you to "move money to protect it."
Amazon sends emails from @amazon.com (or country domains), and every real order notification is visible in Your Orders inside the Amazon app. They never ask for your full card number or SSN by email.
Medicare and Social Security handle most communications by mail. They don't call unexpectedly, don't demand immediate payment, don't threaten to suspend your benefits by phone, and never ask for payment by gift card, wire, or cryptocurrency.
The universal check: Don't act on the message. Open the organization's app or website directly (type the address yourself, don't click the link), or call the number printed on your card or statement. If the message was real, you'll see it there. If it wasn't, you just caught it.
About this tool
This scam checker is built by Ask Safely, a privacy-first AI assistant powered by Anthropic's Claude. It's free, it doesn't require an account, and it doesn't store what you paste.
Ask Safely's full product gives you unlimited private AI chat with auto-deleting conversations. Expert members also get Lookout, a guided scam-checking agent that goes deeper: checking URLs, analyzing patterns across multiple messages, and walking you through what to do if you've already clicked or sent money.