GitHub Repo Scraper — Stars, Forks, Topics & Activity

Scrape GitHub repositories with their full metadata — stars, forks, watchers, open issues, language, topics, licence, size and the created, updated and last-push dates — by search query, by language and star count, by organisation, or by exact repository name.

No API key required for small runs; add a free token for large ones.

What you get

Field	Description
fullName, name, owner, ownerType	Identity, and whether the owner is a User or Organization
description	Repository description
stars, forks, watchers	Popularity
openIssues	Open issues and pull requests
language	Primary language
topics	Topic tags
license	SPDX identifier, e.g. MIT, Apache-2.0
homepage	Project site
isFork, isArchived, isTemplate	Repository status
hasIssues, hasWiki, hasDiscussions	Enabled features
defaultBranch, sizeKb	Branch and checkout size
createdAt, updatedAt, pushedAt	pushedAt is the real activity signal — updatedAt also moves on metadata edits
url, cloneUrl, apiUrl	Links

Example input

{
  "language": "rust",
  "minStars": 10000,
  "sortBy": "stars",
  "maxRepos": 100
}

Find actively maintained projects in a niche:

{
  "topic": "kubernetes",
  "minStars": 500,
  "pushedAfter": "2026-06-01"
}

List everything an organisation owns, or fetch exact repositories:

{ "owner": "vercel", "maxRepos": 500 }
{ "repos": ["facebook/react", "torvalds/linux"] }

Power users can pass any GitHub search string directly:

{ "searchQuery": "language:go stars:>5000 topic:cli archived:false" }

Example output

{
  "fullName": "clash-verge-rev/clash-verge-rev",
  "owner": "clash-verge-rev",
  "ownerType": "Organization",
  "description": "A modern GUI client based on Tauri...",
  "stars": 147543,
  "forks": 10600,
  "openIssues": 474,
  "language": "Rust",
  "topics": ["clash", "clash-meta", "clash-verge", "linux"],
  "license": "GPL-3.0",
  "pushedAt": "2026-09-26T10:01:51Z",
  "isArchived": false,
  "url": "https://github.com/clash-verge-rev/clash-verge-rev"
}

Who uses this

• Developer-tool marketing — find the repositories, and therefore the maintainers and communities, in your category
• Technical recruiting — surface active projects in a language or framework, then their owners
• Open-source strategy — benchmark your project against competitors on stars, issue load and push cadence
• VC and deal sourcing — fast-growing new repos are a leading indicator; combine createdAfter with minStars
• Security and supply chain — inventory an org's repos with licences, archived status and last-push dates
• Dependency and licence audits — license plus isArchived finds abandoned dependencies before they bite

Rate limits — read this for large runs

GitHub allows 60 requests per hour without authentication and 5,000 with a token. Each request returns up to 100 repositories, so an unauthenticated run comfortably handles a few thousand repos per hour, and the actor reports the remaining quota in the log.

For anything bigger, create a free personal access token at github.com → Settings → Developer settings → Personal access tokens and paste it into the GitHub token field. No scopes need to be granted for public repository data. If you do hit the limit without a token, the actor says so explicitly rather than returning a silent partial result.

Two API limits worth knowing

• GitHub's search API never returns more than 1,000 results per query, no matter the page. The actor reports the true total match count so you can see when you are hitting it, and the fix is to narrow the query — split by star band (stars:1000..5000), by language, or by creation year.
• Listing by owner alone bypasses that cap. When you set only owner, the actor uses the repository-listing endpoint rather than search, so it can walk an organisation's entire catalogue.

Pricing

Pay per result. Each repository returned counts as one result.

Notes

• watchers is GitHub's true subscriber count where available; the search API returns a field of the same name that actually mirrors the star count, and the actor prefers the real one.
• license is the SPDX identifier; repositories with a non-standard licence file report the licence name instead, and unlicensed repositories report null.
• This is an unofficial actor and is not affiliated with GitHub.