“Fake profile factory”: how profiles and ads can multiply online

Published by Diogo Godoy · Updated on September 26, 2026

What can be observed — and what still needs proof

When identical names, photos, age, city, description and contact details appear at different URLs, there is a pattern consistent with republication. The comparison can show duplication, but by itself it does not reveal which page came first, who copied it or whether automated collection was used.

A technical analysis assisted by Claude identified repeated matches in images, text and ad structure at scale, consistent with scraper collection and automated republication. This finding strengthens the automation hypothesis in the cases examined. Even so, without access to internal logs, code or infrastructure, it does not by itself establish who operated the system, the intent involved or a corporate relationship between the domains.

What public reports say

There are public reports from people who say they found photos, text or ads published without permission on directories such as SecretHostess, Ladys.one and OnDate. These accounts may indicate a problem worth investigating, but they do not by themselves prove the method used, the source of the material or responsibility for each page.

Erasly does not claim that every profile on these domains is fake, that the services are operationally connected or that every repetition is automated. Models-World, for example, presents itself as an advertising platform and maintains its own edit, contact and removal request pages. This identifies available channels; it does not determine how a specific profile was created.

The name found in the reports reviewed is OnDate, at ondate.com. It should not be confused with other services with similar names.

Tryst, Eros and the common-source hypothesis

In cases reported to Erasly, a person may recognize an ad they originally published on Tryst, Eros or their own website and later find matching elements at other addresses. That timeline matters, but it remains the owner’s account until dates, files and URLs are documented.

The match should be described precisely: which photos are identical, which passages match and which details were changed. Saying one site “automatically clones” another goes beyond what page appearance alone can demonstrate.

How to build useful evidence

Record the full URL, date, time and screenshots showing the page context. Keep the URL you identify as the original and the original files, when available. Compare text passages, phone numbers, usernames and image order.

Reverse image search and file comparison help locate duplicates, but they do not replace timeline analysis. Metadata may have been removed or changed. Domain and hosting information can support an investigation but does not alone prove who published the content.

Removal, delisting and follow-up

The first request is usually directed to the site’s contact, edit or removal channel. Depending on the content and ownership, measures concerning impersonation, privacy, non-consensual intimate content, data protection or copyright may also apply. DMCA is appropriate only when the requester owns or represents the relevant copyright.

Removing a Google result does not erase the source page. Removing one page also does not guarantee that other copies disappear. A replication case therefore needs a URL inventory, priorities, response records and a new check after each measure.

Frequently asked questions