How to check an invoice for AI text and fake documents
Verify a suspect invoice before paying: supplier registry, payment details, text signals and what a detector cannot prove.
Read the guideBlog
Practical guides for scores, documents and review decisions without treating detector output as proof. To see how this detector compares with others, including our own measured recall and false-positive rate, read the AI content detector comparison.
Verify a suspect invoice before paying: supplier registry, payment details, text signals and what a detector cannot prove.
Read the guideCheck a rental listing for AI-generated photos and fake ads: five checks before you pay a deposit or the first month's rent.
Read the guideCheck a damage photo for AI signals with five checks, source material and proportionate follow-up — without accusing anyone of fraud.
Read the guideManual checks for AI-generated photos, plus where a detector does and does not help.
Read the guideHow to recognise Midjourney output, why the newest versions are harder, and what our detector scores on them.
Read the guideHow C2PA Content Credentials work, why they beat a detector score, and where they stop.
Read the guideHow romance scammers use AI profile photos and how to check a match before investing money or trust.
Read the guideA fixed workflow for viral imagery during breaking news, from source to detector signal.
Read the guideCheck a product photo for AI warning signals, reuse and fresh seller evidence before you pay.
Read the guideCheck a PDF for AI writing signals free: extraction quality, per-section review, embedded images and warning-only interpretation. No account needed.
Read the guideCheck a cover letter for AI writing signals without claiming authorship or automatically rejecting candidates. For PDF, DOCX and pasted text.
Read the guideCheck a student essay or assignment for AI signals without automatically accusing the student. Review process evidence, sources and content.
Read the guideCheck an internal or confidential document for AI signals using minimum necessary data, safe extraction and human review.
Read the guideCheck an online review for AI writing patterns and fake-review signals: five checks, what a text detector does and does not show, and the rules for businesses.
Read the guideCheck a job posting for AI photos and fake postings: five checks to verify the company, the images and the channel before you apply or share data.
Read the guideSpot a deepfake advert with a famous face: five checks on the source clip, the offer and the payment route — and why a detector signal is never proof.
Read the guideA payment proof you receive is an image, not a payment. Five checks on status, file, reference and channel — plus what a detector can and cannot see.
Read the guideCheck a suspicious delivery text or email before you click: your own carrier app, tracking number, link check, urgency cues, and never an APK. What a detector can and cannot see.
Read the guideCheck a ticket listing or resale offer before you pay: the official channel, resale protections, sold-out pressure, the seller's account and domain, and what a detector can and cannot see.
Read the guideAn unsolicited job offer on WhatsApp, Telegram or a recorded call? Verify the company outside the offer, never pay to work, and never send an ID copy. Plus what a detector can and cannot see.
Read the guideAn unexpected collection letter about a debt you do not recognise deserves checks, not a payment. Validate the debt and the collector, inspect the pressure tactics, and never pay through a link. What a detector can and cannot see.
Read the guideA message or call from someone you know who urgently needs money, with a voice that sounds right. Verify on your own saved number, use a code word, and recognise the family-emergency payment route. What a detector can and cannot see.
Read the guideYour account is taken over and your contacts receive money requests in your name. Five steps to remove the intruder, warn your contacts and limit the damage — plus five checks for anyone who receives such a message.
Read the guideAn offer that exists only as a social ad, a shop that launched last week and only takes prepayment: five checks before you order — and what a detector does and does not see.
Read the guideSomeone calls claiming to be your bank, says there is fraud on your account, or tells you to move money to a "safe account": five checks, why the number on your screen proves nothing — and what a detector does and does not see.
Read the guideA text or email about an overdue tax debt with a payment link and a threat of seizure: five checks, why the sender proves nothing, and what a detector does and does not see.
Read the guideIn a video call, a face and a voice can be faked in real time. Five checks before you trust someone on screen, what a detector signal does and does not say, and what to do if money has already been sent.
Read the guideA platform that shows returns but never pays out: five checks on licensing, customer service, the payment route and the final "tax or fee" stage — and what to do if you already deposited.
Read the guideBlackmail with intimate images, or with AI-made ones: five steps — pay nothing, preserve evidence, get images removed through Take It Down or StopNCII, report it, and tell someone. And why a detector score is not the decision here.
Read the guide'Easy money' on Snapchat, Telegram or a job offer? Once your account forwards someone else's money, you are the money mule. Five checks, the real consequences (criminal record, repayment, years without banking) and what to do if it already happened.
Read the guideAn urgent, confidential payment request from 'the director'? In CEO fraud the tone, the timing and the project all fit, and the sender address can be faked. Five checks, why dual authorisation is the decision, and what to do if money already moved.
Read the guideNo signal for no reason, a text about an 'expiring SIM' or a caller claiming to be your carrier: how SIM swapping works, what to do immediately, and why an SMS code stops being a second factor. Plus what a detector does and does not see.
Read the guideA text or email saying your payment card was linked to Apple Pay or Google Wallet, with a number to call. How to check it without trusting the sender: your own wallet, your own bank, and never share a verification code. Plus what a detector can and cannot see.
Read the guideA buyer who wants to pay cash and have a courier collect, or a 'paid' screen while the money has not arrived. How to break the structure: your own bank first, no ID copies, no payment links, nothing leaves until the credit is on your account.
Read the guideQR fraud sends you somewhere other than you expect: a sticker pasted over a legitimate code, or a QR code in a fake letter, email or invoice. Check the destination before you pay or enter any details.
Read the guideAn empty box or a different item coming back, or a damage photo that was generated with AI: how to record a return, assess a claim without accusing anyone, and what a detector can and cannot say.
Read the guideA traffic fine that arrives by email or text is always fake. How to check a fine in your own DigiD portal, recognise the email with an attachment and a payment threat, and what a detector can and cannot say.
Read the guideRecognise a fake vacation rental and post-booking phishing: stay inside the platform, verify the address and the "payment verification" messages, and know what a detector can and cannot see.
Read the guideA login page that looks exactly like your bank, your digital ID or a platform: how to check the address before you enter anything, and what to do if you already logged in.
Read the guideA pop-up with a phone number, or someone claiming to be Google or Microsoft asking for your verification code or remote access: five checks before you share anything, and what to do if you already did.
Read the guideA message that you won a prize, voucher or lottery, or a "lottery employee" on the phone: five checks before you respond, and what to do if you already paid or sent an ID copy.
Read the guide