GitHub Search Scraper — Repos, Users & Issues

Search all of GitHub and get clean, structured rows — repositories, users, issues/PRs or code — plus full repo detail lookups by owner/name. Built on GitHub's official REST API with rate-limit backoff. Add a free personal access token to run reliably at 5,000 req/h.

Perfect for developer lead-gen, open-source market research, star/trend tracking, dependency analysis, or feeding a dataset.

What you get

Output depends on searchType (or the repos lookup):

• Repositories — fullName, owner, description, url, homepage, language, stars, forks, watchers, openIssues, topics, license (SPDX), isFork, archived, defaultBranch, createdAt, updatedAt, pushedAt
• Users — login, url, userType, name, company, blog, location, bio, publicRepos, followers, following
• Issues / PRs — number, title, url, state, user, labels, comments, createdAt, updatedAt, closedAt, body (a type of issue or pull_request)
• Code — name, path, repository, url, sha

Every row carries a type field so mixed exports stay unambiguous.

Modes

• Search — pass a query in GitHub search syntax and choose searchType (repositories, users, issues, code). Sort and order supported. GitHub caps search itself at 1,000 results.
• Repo detail — pass repos (e.g. ["facebook/react"]) to fetch full repository objects directly, no search needed.

Input

Field	Type	Default	Description
searchType	string	repositories	repositories, users, issues, code
query	string	—	GitHub search syntax, e.g. stars:>10000 language:rust
sort	string	best match	Repos: stars/forks/updated. Users: followers/repositories/joined. Issues: created/updated/comments
order	string	desc	desc or asc
repos	array	—	Fetch specific repos by owner/name
token	string (secret)	—	GitHub PAT → 5,000 req/h (read-only public_repo scope is enough)
maxResults	integer	100	Cap on results (GitHub search maxes at 1,000)

Provide at least one of query or repos.

Example input

{
  "searchType": "repositories",
  "query": "stars:>50000 language:typescript",
  "sort": "stars",
  "maxResults": 100
}

Devs in a city (lead-gen):

{
  "searchType": "users",
  "query": "location:berlin followers:>500",
  "sort": "followers",
  "maxResults": 50
}

Example output (one repository)

{
  "type": "repository",
  "id": 10270250,
  "fullName": "facebook/react",
  "name": "react",
  "owner": "facebook",
  "description": "The library for web and native user interfaces.",
  "url": "https://github.com/facebook/react",
  "homepage": "https://react.dev",
  "language": "JavaScript",
  "stars": 228000,
  "forks": 46700,
  "watchers": 228000,
  "openIssues": 970,
  "topics": ["javascript", "react", "frontend", "ui"],
  "license": "MIT",
  "isFork": false,
  "archived": false,
  "defaultBranch": "main",
  "createdAt": "2013-05-24T16:15:54Z",
  "updatedAt": "2026-07-11T09:12:00Z",
  "pushedAt": "2026-07-11T08:03:22Z"
}

Why this actor

• Four entities + detail lookups in one actor — repos, users, issues/PRs, code, and direct owner/name fetches, all normalized.
• Full GitHub search syntax — every qualifier works (stars:, language:, location:, repo:, state:, …).
• Reliable at scale — primary and secondary rate-limit backoff; add a token for dependable 5,000 req/h throughput.
• Flat, fair pricing — no credit gating.

Notes

• Token (recommended, BYO): without a token GitHub limits unauthenticated use to ~10 search req/min and lower per-hour caps — fine for small runs, unreliable at scale. A read-only PAT (public_repo scope) lifts you to 5,000 req/h. It's stored as a secret. The actor still runs keyless; it just warns and may throttle.
• GitHub search returns at most 1,000 results per query regardless of maxResults — narrow the query to page deeper.
• Not affiliated with or endorsed by GitHub, Inc.

Pricing

Pay per result — $2.00 per 1,000 results. No subscription or platform fees; you only pay for the results you get.

Related actors

• Hacker News Scraper — stories & comments for dev-community monitoring.
• Stack Exchange Scraper — Q&A across Stack Overflow and sibling sites.
• Package Registry Scraper — npm / PyPI / crates package metadata.