Honest by Default: Why PROBE Says Who It Is

A username search is only as trustworthy as the way it gathers evidence. There are easy shortcuts: pretend to be a web browser, borrow a search engine crawler's identity, or rotate disguises until a page finally answers. PROBE takes none of them.

What PROBE actually sends

Every request identifies itself as PROBE/1.0 (+https://usernameprobe.com): a plain name and a link that any site operator can follow to see who we are and what we do. We do not send the browser fingerprint hints that real browsers attach, because PROBE is not a browser and should not claim to be one.

Older catalog entries that presented a search engine crawler's identity to get a friendlier page have been removed. Where a platform offers an official API that needs a key, the key is attached only over HTTPS and only to that platform's exact host, never forwarded anywhere else.

Why not wear a disguise?

Three reasons. First, it misleads the people who run the sites we check. Second, an answer obtained under a false identity is weaker evidence: pages served to crawlers or challenged visitors often differ from what a person would see. Third, being identifiable keeps us accountable. A platform can recognise PROBE, contact us, or decline to answer.

When a platform declines, we treat that as the answer. A challenge or refusal is reported as blocked, not quietly retried under another name until something gets through. You can read more about why PROBE prefers an honest unknown over a confident guess.

Terms come before coverage

In October we reviewed every manual-only source and every automatic source that had stopped producing decisions. For each one we recorded the technical result alongside the platform's published terms and robots rules.

Several sources where a simple check would technically work, including Pinterest, Poshmark, Scribd, Wellfound and Society6, remain manual because their terms prohibit automated access. Earlier, Figma, Tellonym, eToro, Pr0gramm, HackerOne and Spotify were returned to manual for the same reason. Manual sources still appear in your results with a direct profile link, so you can open the page yourself in seconds.

Being a good neighbour

• One request per source per search. PROBE checks what you asked about, not whole directories.
• Back off when asked. Long Retry-After instructions are honoured, and a source that keeps refusing is paused instead of hammered.
• Bounded capacity. Only a small number of scans run at once; the rest wait in a short queue.
• Nothing kept. Profile details are shown for the current scan only. PROBE does not build a profile database.

What it costs, and what it buys

Honesty costs coverage. Some sources answer only disguised visitors, and some will need paid network infrastructure that we are still evaluating. Others will stay manual for as long as their terms say no. We would rather show unknown or a manual link than an answer gathered under a false name.

What it buys is evidence you can explain. When PROBE reports a result, you know exactly how it was obtained: an identified request, to a public page, within the platform's rules.