Start with one clear face photo
Use one visible face with reasonable lighting and minimal obstruction. Better source photos usually produce stronger candidate matches.
How 221B works
221B is structured as a public-web reverse face search workflow: upload one clear face photo, surface candidate pages, then inspect the original sources before deciding what a match means.
The core promise
Step by step
Use one visible face with reasonable lighting and minimal obstruction. Better source photos usually produce stronger candidate matches.
221B is designed to look across public profiles, articles, blogs, and other crawlable pages where the same face may appear in different photos.
A face match is a prioritisation signal, not a verdict. The source page, username, bio, dates, and surrounding page context decide whether a result is meaningful.
The intended workflow is to compare multiple public clues before treating anything as confirmed. This is what makes the process defensible.
What you can get
What it does not do
Best fit
Useful when you need to see whether the same face appears elsewhere under different names, bios, or stories.
Useful when you want to review public traces before trusting identity claims or meeting someone offline.
Useful when you need public-web discovery plus a manual review process that can stand up to scrutiny.
FAQ
No. Google is strongest at finding duplicate or near-duplicate images. 221B is positioned around finding the same person across different photos on public-web pages.
No. 221B is described as a public-web face search workflow. Private, hidden, or gated content is outside the intended scope.
Open the source page, compare the public identity clues, and decide based on the page context. A strong similarity score should never be treated as automatic proof.
Next step
This page exists to set expectations clearly. The next decision is usually whether you want to buy credits now or read more about scope, limits, and manual-review standards first.