Start with the claim, not the person
A dating contact sends a screenshot or link to a public CFPB complaint page and says the payment service has a clean government record, so you should send money through it. Before you decide anything about that person, decide what the displayed result actually says: which company, which product, which date range, and which filters were visible when the page was captured.
A complaint record is a public-data statement about how complaints were handled in one system. It is not a licence, a safety certificate, or an approval of a specific payment request. The gap between “this page shows few or no complaints” and “send money through this service” is the gap the reader has to check.
What the CFPB database does and does not cover
Three coverage limits matter before you read any count.
Publication eligibility is limited to complaints sent to companies for response; the database does not publish complaints referred to other regulators. Source
That means an empty or small result is partly a statement about which complaints entered the database, not only about the company.
The database is not a statistical sample and does not represent all consumer experiences with a company or product. Source
So a count is not a rate, and a low count is not a clean bill of health.
Complaint volume should be read with company size and market share in mind. Source
A large company with many transactions can accumulate complaints that a smaller one never would, and a small company can look quiet simply because it is small. Neither direction tells you whether a particular payment request is safe.
What changed in August 2026
Some screenshots show complaint narratives or charts that may no longer appear. That change has a dated public explanation.
In its 2026-08-14 announcement, CFPB said it was ceasing publication of complaint narratives and visualizations. Source
The announcement says complaint narratives are unverified and do not necessarily describe a legal violation. Source
It says CFPB would continue collecting, monitoring and responding to consumer complaints. Source
Read together, these statements mean a narrative that is now unavailable is not, by itself, evidence that an incident never happened, was disproved, or was removed because of wrongdoing. An older screenshot is also not automatically fake: it may simply predate the publication change. What you cannot do is treat the absence of a narrative today as a new finding about the company or about your date.
Separate the visible claim from what it supports
The table below keeps each row about the public page, not about the person who sent it. Record only what is visible: company name, product name, date range and any filters shown. Do not ask for or use anyone’s private complaint account, bank records, receipts, IDs or chat exports.
| What the reader sees | Source scope it reflects | What it supports | Remaining question | Next action |
|---|---|---|---|---|
| Empty or very small displayed result | Complaints in this database for the named company/product and visible date range | That few or no matching complaints were published in that scope | Was the company or product name correct, and is the date range the one you care about? | Open the official page yourself and re-check the name and dates |
| A larger complaint count | Same database, same visible filters | That more complaints were published in that scope | How large is the company and its market share, and do the complaints concern the same product you are being asked to use? | Compare the product and period, not just the number |
| An older narrative or visualization screenshot | A page captured before the 2026-08-14 publication change | What the page displayed at that time | Does a current official page still show it, and does it describe this company and product? | Check the current official page before relying on the screenshot |
| A result for a similarly named company or a different product or date range | A different scope than the one in the payment request | Nothing directly about the service you were asked to pay through | Which exact legal or trading name, product and period does your date’s claim refer to? | Ask for the exact name and period, then look it up yourself |
In every row, the bounded inference is the same: the database’s coverage limits mean a count or an absence cannot be turned into a company rating, an approval of a payment request, or a finding of misconduct. That is a conclusion about the source’s scope. It is not a finding by the CFPB about your dating contact, and it is not a finding that the person is lying.
If all you have is a screenshot
You do not need to upload bank records or private messages to check a public page. If you have an ordinary, non-sensitive screenshot of a public CFPB page that you are authorised to review, TraceSeek’s image text search can use the visible company and product wording as search clues. The app is described as reading OCR text in a supplied image to find relevant public content and profile pages; it does not verify complaint counts or certify a service. You still compare the wording it surfaces against the image, then open an accessible original source yourself.
For a broader sweep around the exact organisation and the exact claim, the app’s deep search expands public-source retrieval using AI-assisted search and search engines. It does not guarantee an exhaustive search of the web, and it does not decide whether the payment is safe. The documented username scan, with its free first scan and 20–30 second statement, is a separate input and is not the same as checking a company page.
A practical order: read the visible name and dates on the screenshot, use the wording as search clues, then open the official page directly from a source you already trust rather than from a link in the message. The guide on checking sources in an AI-assisted public web search explains how to keep the original source in view. The app page describes the public-text input route.
What this does not tell you about the payment
A public complaint record says nothing about whether a specific transfer will arrive, whether a recipient is who they say they are, or whether a payment request is legitimate. A username match, a profile photo, or an empty result cannot prove identity. If the only support for “send the money” is a government-record claim that you cannot pin to an exact company, product and date range, you can decline to rely on that claim rather than pay to test it.
If a separate claim is made that the service is authorised by a financial regulator such as the FCA, that is a different check with its own official contact route; treat it as a separate claim, not as something a CFPB complaint page establishes.
The question to ask before you send anything
Which exact company, product, date range and official source does this claim describe, and where does it actually say this payment is approved?
If the answer cannot be supported by a page you can open yourself, the honest conclusion is narrow: the public record does not approve this payment. That is enough to pause, ask for the exact details, and decide on your own terms.