Sherlock Pro — Username Search & Identity Resolution

Search for usernames across 400+ social networks, narrow results down to just the platforms you care about, and — uniquely — figure out which account on a target platform actually belongs to a known profile, even when squatters and lookalike accounts are in the way.

What this actor does

• Search 400+ social networks in a single run, or filter down to just the ones you need via a dropdown of every supported network
• Batch processing — search multiple usernames at once, with {?} wildcard expansion for username variations (e.g. john{?}doe → john_doe, john-doe, john.doe)
• Identity resolution — give it a known account on one platform and a target network, and it either auto-discovers or verifies candidate accounts there, ranking each by a transparent, multi-signal confidence score (bio backlink, display-name/avatar similarity, verified badge, follower count, username similarity) — never a black-box match
• Honest "no match" reporting — if no candidate clears the confidence bar, the actor says so instead of guessing
• Empty fields are omitted — no null, no placeholders; every field present is real data
• No login required — works entirely with publicly available information

Output

Broad Search (mode = broadSearch)

• username — the username that was searched
• links[] — URLs of profiles where the username was found (omitted entirely if none found)

Identity Resolution (mode = identityResolution)

One record per target network:

• sourceProfile — the known account being resolved against, with any enrichment found: platform, handle, profileUrl, displayName, verified
• targetNetwork — the network this record's candidates were evaluated on
• candidateMode — discover or verify
• candidatesEvaluated — how many candidates were scored
• bestMatch — the highest-confidence candidate (handle, profileUrl, confidenceScore, confidenceLabel, signals) — omitted entirely if none clears the confidence bar
• candidates[] — every scored candidate, ranked by confidence, each with its own signals breakdown
• warnings[] — non-fatal notes (e.g. a signal couldn't be computed, or auto-discovery isn't supported for this network) — omitted if empty

Input

Field	Type	Default	Description
mode	string	broadSearch	broadSearch or identityResolution
usernames	array	["johndoe"]	Usernames to search (mode=broadSearch). Supports {?} wildcard.
targetNetworks	array	["GitHub"]	Networks to limit the search/resolution to — dropdown of every supported network. Optional filter in broadSearch (omit to search all 400+); required (1+) in identityResolution; exactly one required when candidateMode is verify.
sourceProfile	object	{"platform": "instagram", "handle": "xyz"}	The known account to resolve against — {platform, handle} (mode=identityResolution)
sourceOverrides	object	–	Manual displayName/bio/avatarUrl for the source, used when the source platform can't be auto-read or to supplement missing data (mode=identityResolution)
candidateMode	string	discover	discover — auto-find candidates via search + username-variation heuristics; verify — score only the handles you supply (mode=identityResolution)
candidateHandles	array	–	Candidate handles to score (required when candidateMode is verify)
maxCandidatesToDiscover	integer	20	Caps how many candidates discover mode generates/searches (1–60)

Example: filter Broad Search to specific networks

{
  "mode": "broadSearch",
  "usernames": ["johndoe"],
  "targetNetworks": ["GitHub", "YouTube"]
}

Example: discover the real account on a target network

{
  "mode": "identityResolution",
  "sourceProfile": { "platform": "instagram", "handle": "xyz" },
  "targetNetworks": ["GitHub"],
  "candidateMode": "discover",
  "maxCandidatesToDiscover": 20
}

Example: verify a shortlist of suspects

{
  "mode": "identityResolution",
  "sourceProfile": { "platform": "github", "handle": "torvalds" },
  "targetNetworks": ["TikTok"],
  "candidateMode": "verify",
  "candidateHandles": ["torvalds", "torvalds_official", "real_torvalds"]
}

Example: batch username search with wildcard expansion

{
  "mode": "broadSearch",
  "usernames": ["john{?}doe", "janedoe"],
  "targetNetworks": ["GitHub", "Reddit", "Twitter"]
}

Use Cases

• Influencer verification — confirm which TikTok/Instagram/Twitter account genuinely belongs to a known creator before running a partnership or ad spend
• Brand protection — find squatter and impersonator accounts across networks and see exactly why they were flagged as low-confidence
• OSINT investigations — discover all online accounts associated with a username, filtered to the networks that matter for the case
• Digital forensics — cross-reference identity signals across platforms with a transparent, auditable trail
• Recruiting / vendor vetting — cross-check a candidate's or vendor's professional and social presence before engaging

FAQ

Can I limit results to just the networks I care about?

Yes. Set targetNetworks in broadSearch mode (e.g. ["TikTok", "Instagram"]) — pick from the dropdown of every supported network — and only those sites are checked, so you don't pay for or wait on the other 400+. identityResolution mode additionally accepts Reddit as a target since it resolves candidates through Reddit's own public search API rather than Sherlock's site list.

How does it tell the real account apart from fake/lookalike accounts?

This is what identityResolution mode is for. Given a known source profile, it finds candidate accounts on the target network (either by searching the platform directly or by generating and existence-checking common username variants like numeric suffixes and "official"/"real" patterns), then scores every candidate on independent signals: whether the candidate's bio links back to the source account, how similar the display names are, how similar the profile pictures are, verified-badge status, relative follower count, and username similarity (case-insensitive). The scores are weighted so that bio backlinks and display-name/avatar similarity — signals fakes rarely have — count far more than raw username similarity, which is exactly what squatter accounts (xyz0, xyz1, xyz2) try to exploit. A bio backlink only ever helps a candidate when found — a real profile whose bio simply doesn't mention the source platform (common; most people don't cross-link everywhere) is never penalized for it. The full signal breakdown is always included so you can see exactly why a candidate was or wasn't picked. If no candidate is confidently the real account, the actor reports that honestly instead of guessing.

What if I already know the suspect accounts and just want them ranked?

Use candidateMode: "verify" with candidateHandles set to your list of suspects and a single targetNetworks entry. The actor skips discovery and scores exactly those candidates.

Why is auto-discovery unavailable for some networks?

TikTok's public existence-check reports every handle — including random strings — as "existing," so auto-generated candidates there would be fabricated rather than real. Auto-discovery (candidateMode: "discover") is disabled specifically for TikTok for this reason; candidateMode: "verify" still works fully, with the same transparent scoring, once you supply your own candidate list.

How many social networks does it search?

Broad Search covers the same 400+ platforms as the original Sherlock — GitHub, Twitter/X, Instagram, Reddit, TikTok, LinkedIn, YouTube, and hundreds more.

Does it find private or deleted accounts?

No. Only publicly accessible profiles are checked — private, deleted, or restricted-visibility accounts won't appear.

What does the {?} wildcard do?

Expands a username into three variations using common separators: _, -, and .. For example, john{?}doe searches john_doe, john-doe, and john.doe.

Does this require a login or API key?

No. Everything works with publicly accessible information — no authentication, cookies, or API keys required.

Can false positives occur?

Yes, occasionally — on both Broad Search (a site may report a username as existing when the page is actually an error page) and Identity Resolution (signals can be sparse for platforms that block scraping, e.g. Twitter/X, lowering confidence). Always review the signals breakdown before treating a match as certain.

Is this affiliated with any of the platforms it searches?

No. This is an independent, third-party actor using only publicly accessible information; it is not affiliated with GitHub, Reddit, TikTok, Instagram, Twitter/X, LinkedIn, YouTube, or any other platform it searches.

What output formats are available?

Results can be exported as JSON, CSV, Excel (XLSX), HTML, RSS, or XML directly from the Apify platform.