What Tracehound is, and how a search works

Tracehound answers one question: where else does this identifier show up? You give it a username, an email address or a phone number. It asks a long list of public sources, all at once, whether they know that identifier, and shows you every match as it arrives, with the profile details each source makes public.

That is the whole product in a sentence. This post explains what happens in the seconds between pressing Search and seeing the last result, what you see on screen, and just as importantly, what Tracehound deliberately does not do.

The short version

• Three search types: username, email and phone.
• Every search runs live against public sources, in parallel. There is no stored database of people behind it.
• Results stream in source by source, so the first matches appear while slower sites are still answering.
• Result data is processed in memory and never written to our database.
• A match is a research lead, not proof of who someone is.

What happens when you press Search

1. The identifier is cleaned up

Small differences in how people type an identifier should not change the answer. A leading @ is removed from a username. An email address is lower-cased. A phone number is parsed and converted to the international E.164 format, which also tells us early whether the number is valid at all. If the input cannot be a valid identifier of the chosen type, the search stops there and costs nothing.

2. The opt-out list is checked

Anyone can block searches for their email address by confirming a one-time code, and can ask us to block a username or phone number. Before any source is contacted, the identifier is compared against that list. Opted-out identifiers are stored only as keyed one-way hashes, so the comparison works without the list holding readable addresses or numbers.

3. Every source is asked at the same time

Each supported source has a small module that knows how that site reveals whether an account exists: a profile URL that loads or does not, a public API, or a page that says "user not found". Tracehound runs all relevant modules in parallel. A search for a username fans out to hundreds of sites; an email or phone search goes to the smaller set of sources that support those lookups.

4. Each source answers in one of three ways

This is the part that matters most for trusting the results. A source can say the account exists, say it does not, or fail to give a trustworthy answer, for example because it rate-limited the request or returned a page we do not recognize. That third state is reported as no answer, not quietly folded into "not found". Guessing would be faster to build and worse to use.

5. Matches stream to your browser

As soon as a source confirms an account, the result is sent to your screen. You do not wait for the slowest site before you can start reading. A progress counter shows how many sources have answered so far. A full search typically finishes well under a minute after you press Search.

What a result shows

Each match is a profile card. What is on it depends on what the source itself publishes, and every field is taken from that source as-is. Nothing is inferred or filled in.

• On every plan: the site, a link to the profile, display name, handle, avatar, bio, and follower, following and post counts where available.
• On Pro and Max: extended fields such as location, website, join date, last active date, other accounts the profile links to, and platform details like ratings or repository counts.

You can switch between cards and a compact table, filter by category (social, developer, gaming and so on), sort, and search within the results. On Pro and Max you can export the set as CSV or JSON.

Leads and Deep Search

Profiles often mention other identifiers in public: a GitHub profile lists an email address, a Mastodon bio names an older handle, a Keybase page links several accounts. On Pro and Max, Tracehound collects these as leads and shows them in a strip above the results. Each one is a one-click follow-up search.

On Max, Deep Search follows the leads for you. It searches the new identifiers it finds, then the identifiers those results reveal, up to 3 layers deep and up to 5 identifiers per layer. Each layer costs one search, and a layer that ends up with nothing new to follow is refunded. Every result is tagged with the identifier that led to it, so you can always trace why it is there.

What we keep

If your search history is switched on, we store the query, its type, the number of results, the status and the time. That is enough to find and rerun a past search. The results themselves are not stored. You can delete single history entries, clear all of them, or turn history off. Signed-out demo searches have no history at all.

What Tracehound is not

It is not a people database. There is no profile of anyone sitting on our servers waiting to be looked up, and nothing to buy, sell or leak.

It is not a background check. Tracehound is not a consumer reporting agency, and it must not be used for decisions about employment, housing, credit or insurance. Our guide to responsible OSINT explains why.

It is not proof of identity. A match tells you that an account with that exact identifier exists on that site. Whether it belongs to the person you have in mind is a separate question, and one you answer by comparing details. Our post on reading username results critically covers how.

Trying it

Without an account you can run 3 searches a day. The first 8 matches are shown in full and the rest are listed by site name. A free account gives you 25 searches a month with every result visible. Pro ($19 a month) adds extended fields, leads and export; Max ($59) adds Deep Search and a much larger allowance. One search is one token, whatever the type, and a search that fails is refunded automatically.

The best first search is your own handle. It shows you exactly what the tool does, and it tells you something useful about your own footprint at the same time. If you want to go further, our footprint audit guide turns that first search into a clean-up plan.