GitHub Candidate Scorer

Score developers against a job by reading their ACTUAL code — not just keywords on a profile.

Give it a job description and either a list of GitHub usernames or a search query. For each developer, it reads their real profile, repositories, languages, topics and README content, then scores how well they fit the role — with evidence (which projects and skills justify the score), strengths, gaps, an estimated seniority, and recovered contact details.

This is what a great technical screener does — read the real work — done at scale. Uses the open GitHub API only (no scraping, no legal grey area).

Quick start

1. Fill in the input — set the fields in the input tab (prefilled with a working example).
2. Click Start — the Actor runs and writes one record per item to the dataset.
3. Get your data — export to JSON, CSV or Excel, or pull it via the API.

// input
{ "jobDescription": "Senior Rust engineer, systems/CLI, open-source track record", "usernames": ["someuser", "another"] }

// per-candidate output
{
  "username": "someuser", "name": "Jane Dev", "matchScore": 88, "seniority": "senior",
  "topLanguages": ["Rust", "Go", "TypeScript"],
  "matchReason": "Strong systems background: authored a popular Rust CLI and contributes to async runtimes...",
  "strengths": ["2.5k-star Rust CLI", "Active OSS maintainer", "Systems-level projects"],
  "gaps": ["No obvious distributed-systems work"],
  "evidence": ["repo: fast-grep (Rust, ★2.5k)", "repo: tokio-utils contributions"],
  "location": "Berlin, DE", "email": "jane@example.com", "profileUrl": "https://github.com/someuser"
}

Use cases

• Technical sourcing / recruiting — rank a shortlist (or a search) of developers against a specific role, with evidence.
• Hiring triage — turn a pile of GitHub profiles into a ranked, explained candidate list.
• Agent tool — via the Apify MCP server, an agent can score a developer against a role on demand.

Input

Field	Type	Default	Description
jobDescription	string	— (required)	The role to score against.
usernames	string[]	—	Specific GitHub usernames to score.
searchQuery	string	—	GitHub user-search query to find candidates (used if no usernames).
maxResults	integer	25	Max candidates to score (one charge each).
model	string	gpt-4o-mini	OpenAI model.

Output

One record per candidate: matchScore, seniority, matchReason, strengths[], gaps[], evidence[], topLanguages[], name, location, email, blog, followers, publicRepos, profileUrl. Top candidates in the run SUMMARY.

Pricing

Pay-per-event: one candidate-scored event per developer scored.

Environment variables

Variable	Required	Purpose
OPENAI_API_KEY	Yes	Powers the scoring.
GITHUB_TOKEN	Recommended	Free GitHub personal access token — raises the GitHub API limit from 60 to 5,000 requests/hour. Without it, only a few candidates per run.

Notes

• Reads profile + repositories + top READMEs (real "what did they build" signal). Forks and archived repos are discounted.
• Scoring is evidence-based; provide a detailed job description for best results.

License

ISC.

Related actors

• Nordic Tender Matcher (AI) — AI agent that scores Norwegian, Finnish and Swedish public tenders against your company
• EU & Nordic Grant Finder (AI-matched) — Find open EU + Nordic public-funding calls matched to your company with AI — eligibility
• AI SEO + AEO/GEO Site Auditor — Crawl & Auto-Fix — Crawl a whole site and get an AI-prioritized technical + content + AEO/GEO audit with