GitHub Repo Stats & Enrichment

Enrich GitHub repos or a search query with stars, forks, activity, license, languages, and top contributors.

Give the Actor a list of repositories or a search query. It returns one clean, structured record per GitHub repository so you can compare projects, track activity, or feed live stats into your own tools.

What it does

• Search GitHub repositories by query, ranked by stars, or pass an explicit list of owner/repo names.
• Per repo: stars, forks, watchers, open issues, primary language, license, topics, and archived status.
• Repo age in days, creation date, and last push date.
• Optional language byte breakdown for each repository.
• Optional top contributors (login and contribution count) for each repository.
• Aggregate summary (repos returned, total stars, most-starred repo) in the run's OUTPUT.

Example input

Search for the top LLM agent frameworks:

{
  "query": "llm agent framework",
  "limit": 5,
  "includeContributors": true,
  "includeLanguages": true
}

Enrich a specific list of repositories:

{
  "repos": [
    "apify/crawlee",
    "langchain-ai/langchain"
  ]
}

Input

Field	Type	Description
query	string	GitHub search query. Used when no repos list is given.
repos	array	Explicit list of owner/repo strings. If set, query is ignored.
limit	integer	How many repositories to return in search mode (default 10, max 100).
includeContributors	boolean	Fetch top contributors per repo (default true).
includeLanguages	boolean	Fetch the language byte breakdown per repo (default true).

Output

Each dataset item is one repository:

{
  "fullName": "langchain-ai/langchain",
  "description": "Build context-aware reasoning applications",
  "url": "https://github.com/langchain-ai/langchain",
  "stars": 92000,
  "forks": 15000,
  "watchers": 620,
  "openIssues": 280,
  "primaryLanguage": "Python",
  "languages": {
    "Python": 12345678,
    "MDX": 234567
  },
  "license": "MIT",
  "topics": [
    "llm",
    "agents",
    "python"
  ],
  "createdAt": "2022-10-17T02:58:36Z",
  "pushedAt": "2026-07-03T18:22:10Z",
  "ageDays": 991,
  "isArchived": false,
  "topContributors": [
    {
      "login": "hwchase17",
      "contributions": 1200
    }
  ]
}

Run summary (OUTPUT)

The run's default key-value store record OUTPUT holds an aggregate for the whole result set:

{
  "mode": "search",
  "query": "llm agent framework",
  "reposReturned": 5,
  "totalStars": 210000,
  "mostStarredRepo": {
    "fullName": "langchain-ai/langchain",
    "stars": 92000,
    "url": "https://github.com/langchain-ai/langchain"
  },
  "generatedFrom": "https://api.github.com"
}

How it works

The Actor calls the public GitHub REST API. In search mode it queries search/repositories sorted by stars. In repos mode it fetches repos/{owner}/{repo} for each entry. For each repository it optionally adds the language breakdown and top contributors, then pushes a clean record to the dataset and writes the aggregate to OUTPUT.

It runs unauthenticated by default, which GitHub rate limits to about 60 core requests per hour and 10 search requests per minute. Set a GITHUB_TOKEN environment variable to raise those limits. Rate-limit responses are handled gracefully with backoff and retries.