Candidate Cross-Platform Resolver

Verify a candidate's real footprint before you reach out — not another "here's a maybe, go check it yourself" name-search tool. Give it any subset of full name, company, LinkedIn URL, GitHub username, personal site, or X/Twitter handle, and it cross-references 12 platforms into confidence-scored matches, each backed by a stated reason, not a guess.

What you get, per lead:

• A confidence-scored match per platform — a numeric score (0.3-0.95) and tier (declared / site-crosslink / correlated / weak) for every hit, so you know how much to trust it before you act on it.
• The actual evidence, in plain English — e.g. "GitHub's own social_accounts field links this LinkedIn URL — the candidate declared it themselves" — so you can judge the match yourself, not trust a black box.
• Zero LinkedIn/X scraping, ever — those only surface when independently corroborated by an official API or the candidate's own site, keeping you clear of the ToS risk that comes with scraping either platform directly.

Sample output (real shape, one entry per lead):

{
  "inputLead": { "githubUsername": "addyosmani" },
  "matches": [
    {
      "platform": "github",
      "url": "https://github.com/addyosmani",
      "handle": "addyosmani",
      "confidenceScore": 0.95,
      "confidenceTier": "declared",
      "evidence": ["GitHub username supplied directly in the lead"]
    },
    {
      "platform": "linkedin",
      "url": "https://linkedin.com/in/addyosmani",
      "handle": null,
      "confidenceScore": 0.95,
      "confidenceTier": "declared",
      "evidence": [
        "GitHub's own social_accounts field links this linkedin URL — the candidate declared it themselves"
      ]
    }
  ],
  "overallConfidence": 0.95,
  "checkedAt": "2026-09-21T10:00:00.000Z"
}

• B2B/recruiting-only. No consumer platforms (Instagram, TikTok, Facebook). LinkedIn and X are never scraped — they only appear as input anchors you already have, or as corroboration when GitHub's own API or the candidate's personal site independently confirms the same link.
• No web scraping of the platforms that matter most for identity — GitHub is an official, authenticated-optional REST API; the only HTML fetched is the candidate's own personal site, for finding self-published links.
• Stateless. Each run resolves the leads you send it — no recurring monitor, no dedupe state. Full behaviour is in ./SPEC.md.
• Part of the RecruitOps suite alongside ats-hiring-signal — use that one to find which target companies are hiring, then this one to verify a candidate before you pitch them.

---

1. Quick start (Apify Console)

1. Open the Actor, paste this into Input, and click Start:

   {
     "leads": [
       { "githubUsername": "addyosmani" },
       { "fullName": "Jane Doe", "company": "Acme Corp" }
     ]
   }
   

2. Results appear in the run's Dataset: one row per lead, each with a matches[] list — platform, URL, a confidence tier (declared / site-crosslink / correlated / weak), a numeric score, and an evidence[] array explaining why each match was made.

3. Strongly recommended: supply a free githubToken (a GitHub personal access token, no scopes needed) — GitHub's unauthenticated rate limit is 60 requests/hour (10/min for name search), which more than a handful of leads per run will exhaust. A token raises this to 5,000/hour.

2. Which fields to give it, depending on what you start with

Each lead needs at least one of these three anchor patterns — pick based on what you already have:

You have	Give it	What happens
Their GitHub username	githubUsername	Most direct — fetches the profile and its declared links straight away.
Their LinkedIn profile	fullName + company + linkedinUrl	Searches GitHub by name (narrowed by company), then checks whether the GitHub profile it finds independently confirms the same LinkedIn URL. If it does, that's a real declared match — two sources agreeing, not just your say-so.
Only a name	fullName + company	Same GitHub search, but nothing to corroborate a LinkedIn/X link against — expect correlated/weak tier results, not declared.

A bare linkedinUrl with nothing else does very little. Since LinkedIn is never scraped (see below), the Actor has no name to search GitHub with — it just echoes your URL back at weak confidence, unconfirmed. Always pair a LinkedIn URL with the name (and ideally company) from that same profile.

3. What a confidence tier means

Tier	Meaning	Score
declared	The platform's own API says the candidate linked this account themselves	0.95
site-crosslink	Found by reading the candidate's own personal site/blog	0.85
correlated	Same handle exists elsewhere, with some corroborating signal	0.6
weak	Name-only or unconfirmed input, no independent corroboration	0.3

A "weak" match is a lead to verify, not a confirmed identity — say so plainly to your own users if you build on top of this.

4. Why this exists

Cheap commodity "find social profiles by name" tools (as low as $2/1,000 results) tell you that they found something and ask you to verify it yourself manually. Enterprise sourcing platforms (SeekOut, hireEZ) charge $150-500+/seat/month, sales-led, for roughly this job. This sits in between: self-serve, pay-per-lookup, and — unlike the cheap tools — every match ships with the actual evidence, not just a guess. See ./SPEC.md §1 for the full reasoning and pricing.

5. Local development

npm ci
npm run check   # typecheck + lint + unit tests (all pure, no network)
npm run dev     # runs against the real GitHub API using local input
npm run build   # compiles to dist/

6. License

Proprietary — see ./LICENSE. © 2026 BuildersArk LLC.