GitHub Scraper API: Repos, Issues, PRs & Users
Pricing
from $5.00 / 1,000 result returneds
GitHub Scraper API: Repos, Issues, PRs & Users
Scrape GitHub repos, issues, PRs, commits, releases, READMEs, users and trending in 12 modes. Pass lists of repos or users, search users, resume after rate limits. MCP-ready, flat JSON for AI agents.
Pricing
from $5.00 / 1,000 result returneds
Rating
0.0
(0)
Developer
Khadin Akbar
Maintained by CommunityActor stats
0
Bookmarked
67
Total users
24
Monthly active users
8 days ago
Last modified
Categories
Features
Share
What is GitHub Scraper?
GitHub Scraper (repos, issues, PRs, commits, releases, READMEs, users and trending) is an Apify Actor for public GitHub data, built for developers, researchers and AI agents. Give it one repository, user or search query, or a whole list of up to 500 repositories or users, and it returns one flat JSON record per result with stars, forks, labels, authors, dates and URLs. No GitHub login of your own is needed.
What's new (12 Nov 2026): list inputs (repos, users), a new user-search mode, faster big runs (rows saved in batches, parallel lookups), a 60-second request timeout with retries, and a clear PARTIAL result with a resume.startPage if GitHub's rate limit is reached. Code search is now available, and big contributor and user lists are looked up in batches, so large runs hit GitHub's rate limit much later.
It accepts one GitHub surface per run, such as repositories, a search query, user logins, or a trending filter. Each record includes mode, type, url, and scrapedAt, plus mode-specific GitHub fields that support repo analysis, issue review, PR inspection, code search, contributor lookup, release tracking, README retrieval, commit history review, user profiling, and trending exploration.
Best fit and connected workflows
This Actor fits workflows that need one GitHub scraping tool with multiple routes:
- Repository research: use
repowhen a workflow starts from one public repository and needs a full metadata snapshot. - Issue and PR review: use
issuesorprswhen a workflow starts from a repo and needs lists, labels, comments, reviews, or review threads. - Code discovery: use
code-searchwhen a workflow starts from a text query and needs code matches across GitHub. - Contributor and release analysis: use
contributorsorreleaseswhen the workflow needs contributor profiles or release history from a repo. - Documentation extraction: use
readmewhen the workflow needs the repository README as markdown and rendered text. - Commit inspection: use
commitswhen the workflow needs commit history, and optionally file diffs. - Profile review: use
userwhen the workflow starts from a GitHub login and needs the profile plus repositories. - Trend scanning: use
trendingwhen the workflow needs trending repositories by language and timeframe.
Practical scenario
A developer advocate named Mina starts with apify/actors-mcp-server and wants to prepare a short update for her team. She runs the Actor in prs mode with state: "closed" and includeReviews: true. The returned dataset record includes the PR url, type: "pr", and scrapedAt, plus PR-specific fields for review state and review comments. Mina uses the review state and PR URL to pick one merged change for a deeper read, then shares the PR link in her team note.
Input fields
| Field | Type | Used in | Description |
|---|---|---|---|
mode | string | all modes | Selects one GitHub surface to scrape. One mode per run. |
repos | array | repo, issues, prs, contributors, releases, readme, commits | List of repositories (owner/name or URL), up to 500 per run, one start fee. |
users | array | user | List of GitHub logins, up to 500 per run. |
maxResultsPerTarget | integer | list runs | Optional cap per repo or user inside a list run. |
startPage | integer | paged modes | Resume helper: set to resume.startPage from a PARTIAL run's OUTPUT. |
repo | string | repo, issues, prs, contributors, releases, readme, commits | GitHub repository in owner/name format, or a full GitHub repo URL. |
query | string | repo-search, code-search, user-search | GitHub search query with qualifiers such as language:, stars:, org:, path:, extension:, and for users location: or followers:>100. |
user | string | user | GitHub user or organization login. |
language | string | trending | Language filter for trending mode. |
timeframe | string | trending | Trending window: daily, weekly, or monthly. |
state | string | issues, prs | Issue or PR state filter: open, closed, or all. |
since | string | issues, prs, commits | ISO 8601 date for returning items updated or created at or after the given point. |
maxResults | integer | all modes | Maximum number of records returned in total across the run. |
includeComments | boolean | issues, prs | Includes comments for each issue or PR. |
includeReviews | boolean | prs | Includes reviews and review comments for each PR. |
includeFiles | boolean | commits | Includes file diffs for each commit. |
Focused JSON input example
{"mode": "prs","repo": "vercel/next.js","state": "closed","includeReviews": true,"since": "2026-01-01","maxResults": 25}
Output fields
The dataset returns one flat record per result. Every record has mode, type, url and scrapedAt; the other fields depend on the mode.
Mode → type | Main fields |
|---|---|
repo, repo-search → repo | fullName, description, homepage, language, topics, stars, forks, watchers, openIssues, license, archived, defaultBranch, createdAt, pushedAt, languages, latestRelease |
issues → issue | repo, number, title, state, stateReason, author, assignees, labels, milestone, commentCount, comments (with includeComments), body, createdAt, closedAt |
prs → pr | repo, number, title, state, draft, merged, mergedAt, author, reviewers, labels, head, base, additions, deletions, changedFiles, commits, reviewCommentCount, reviews (with includeReviews), comments |
commits → commit | repo, sha, message, author, committedAt, parents, verified, stats, files (with includeFiles) |
code-search → code-match | repository, file path and match URL for each code hit |
contributors, releases, readme, user, trending | contributor profiles, release notes and assets, README markdown and text, user profile plus repositories, trending repositories |
A missing repository returns one type: "not-found" record with a _note explaining what to fix, instead of an empty run.
Example record (a real repo mode result, trimmed)
{"mode": "repo","type": "repo","owner": "react","name": "react","fullName": "react/react","description": "The library for web and native user interfaces.","url": "https://github.com/react/react","homepage": "https://react.dev","language": "JavaScript","topics": ["declarative","frontend","javascript","library","react","ui"],"stars": 250881,"forks": 51430,"watchers": 6604,"openIssues": 1411,"defaultBranch": "main","license": "MIT","archived": false,"createdAt": "2013-05-24T16:15:54Z","pushedAt": "2026-10-02T21:13:55Z","languageBreakdown": ["JavaScript","Rust","TypeScript"],"latestRelease": {"tagName": "v19.3.0","name": "19.3.0 (September 9, 2026)","publishedAt": "2026-09-09T18:01:33Z","url": "https://github.com/react/react/releases/tag/v19.3.0"},"scrapedAt": "2026-10-04T15:13:55.579Z"}
How it works
This Actor uses GitHub's public REST API for the selected mode, plus the public GitHub trending page for trending mode. The input schema defines 12 modes: repo, repo-search, issues, prs, code-search, contributors, releases, readme, commits, user, user-search, and trending. Lookups for each page of results run in parallel and rows are saved in batches. If GitHub's rate limit is reached the run keeps what it already delivered, ends PARTIAL, and OUTPUT shows resume.startPage. Repository-based modes normalize owner/name input, and full repository URLs are accepted for the repo field. The output dataset stays flat so downstream tools and agents can read it easily. OUTPUT and RUN_SUMMARY in key-value storage state whether the run was complete, partial, valid but empty, invalid, or failed upstream.
Pricing
This Actor uses Apify Pay per event. Apify platform usage is included in the event price, so you pay only for the events below. Open the Pricing tab in Apify Console for the current event prices.
| Event | When it is charged |
|---|---|
apify-actor-start | Automatically at run start, including runs that return no useful result or fail after starting. The event count depends on the memory selected for the run. |
result | Each useful simple record persisted to the dataset: repos, issues or PRs without comments/reviews, commits without file diffs, contributors, releases, README, users, user-search results, trending repos. |
deep-result | Each useful heavier record persisted to the dataset: code-search matches, issues with comments, PRs with comments or reviews, commits with file details. |
For example, a run that returns twenty simple results is billed for one start event plus twenty result events at the default memory setting. The same twenty PRs fetched with includeReviews: true are twenty deep-result events. Cost example at the rates listed today (check the Pricing tab, rates can change): 1,000 simple results = 1 start event + 1,000 result events, so about 1,000 x the result price plus one start fee. A list of 100 repositories in repo mode is one run, one start fee and 100 results. A valid empty search, invalid input, or a GitHub rate limit can still incur the start charge but no result charge. A run spending limit controls the maximum event charges; check the Pricing tab for its current minimum. Platform usage is included, so there is no separate usage charge.
Use with AI agents (MCP)
This Actor is Apify MCP-ready and usable through Apify MCP as khadinakbar/github-deep-scraper. It is a tool for retrieving structured GitHub data by mode, so an agent can request exactly the surface it needs and then read the dataset rows returned by the run.
Tool description: fetch public GitHub records for a chosen mode, then read the resulting dataset items and OUTPUT / RUN_SUMMARY run outcome records.
Fetch the closed pull requests for
vercel/next.jssince the start of the year, include reviews, and return the dataset items with their GitHub URLs and scrape timestamps.
When interpreting results, treat useful dataset rows as GitHub entities or search results; a diagnostic row reports a missing or invalid input/target and has no result charge. The shared fields identify the source mode, record type, canonical URL, and fetch time. For broader runs, use maxResults to bound how many rows the agent receives. For heavier modes and options, such as code search, PR reviews, PR comments, and commit file details, the deep-result event rate applies. Inspect OUTPUT or RUN_SUMMARY for the terminal outcome and charge counts.
Apify API example
import { ApifyClient } from "apify-client";const client = new ApifyClient({token: process.env.APIFY_TOKEN,});const run = await client.actor("khadinakbar/github-deep-scraper").call({mode: "repo",repo: "facebook/react",maxResults: 1,});const { items } = await client.dataset(run.defaultDatasetId).listItems();console.log(items);
Best results and outcome guidance
Use repo (or repos for a list) when the workflow starts from repositories and needs a broad snapshot. Use user-search to find developers by location, language or followers. Use repo-search when the workflow starts from a topic, stack, or qualifier query. Use issues and prs when the workflow needs discussion context, labels, reviews, or thread comments. Use commits when the workflow needs change history, and enable file diffs when reasoning about touched files matters. Use trending when the workflow needs a current list of popular repositories by language and timeframe. Set maxResults to the smallest number that covers the task so the run stays focused.
Continue the workflow
- Then use Google Play Store All-in-One Scraper — Apps, Reviews, Charts to extend GitHub Deep Scraper — Repos, Issues, PRs, Code Search, Commits with a neighboring software ecosystem research source when the brief calls for Google Play data.
- Then use Shopify App Store Scraper — Apps, Pricing & Reviews to extend GitHub Deep Scraper with ecosystem research on the Shopify side when the brief calls for app-store data.
- Then use App Store Reviews Scraper - iOS & macOS Ratings to extend GitHub Deep Scraper — Repos, Issues, PRs, Code Search, Commits with a neighboring software ecosystem research source when the brief calls for App Store data.
Design note
I found that the dataset contract is intentionally flat: every record includes mode, type, url, and scrapedAt, while the rest of the fields vary by mode.
FAQ
Which GitHub surface should I use for a full repository snapshot?
Use repo for one repository's full metadata.
Which mode is suited to finding repositories by topic or qualifier?
Use repo-search and pass a GitHub search query in the query field.
Which mode should I use for issue threads or PR conversations?
Use issues or prs, and turn on includeComments when you want thread context.
How do I inspect code across GitHub?
Use code-search with a search query. GitHub only allows code search for signed-in callers, so this mode uses the Actor's built-in access token (no setup needed on your side).
Can I scrape many repositories in one run?
Yes. Put them in repos (up to 500, owner/name or URL) and set maxResults for the total; you pay one start fee.
What if GitHub's rate limit stops a run?
The run saves what it has, ends PARTIAL, and OUTPUT lists resume.startPage; run again with that startPage to continue.
How do I find GitHub users by location or language?
Use user-search with a query like location:berlin language:rust followers:>100.
How do I get a repository README?
Use readme with the repository name in owner/name format.
How do I get commit history with file diffs?
Use commits and set includeFiles: true when file-level reasoning matters.
How do I get trending repositories for a language?
Use trending with language and timeframe.
Responsible use
Use this Actor for public GitHub data and keep your usage aligned with GitHub's terms and applicable data-protection rules. Review the live Pricing tab before larger runs, especially when using result-heavy modes or options that add comments, reviews, or file diffs.