GitHub Trending Scraper — Repos & Developers · Daily Feed

Why use this scraper

GitHub's own trending page has no public API. If you want to track which repositories or developers are gaining stars right now, whether for a newsletter, a dev-intel dashboard, an investment signal, or a weekly "what's hot in open source" post, you either refresh the page by hand or build and maintain your own scraper. GitHub changes the page's markup periodically, and because there is no official endpoint for it, anything built in-house tends to break silently the next time GitHub ships a front-end update, and nobody notices until the numbers stop making sense.

This actor reads the real github.com/trending HTML, both the repository leaderboard and the developer leaderboard, and returns structured records: repo name, owner, description, primary language, total stars, stars gained in the selected window, forks, and the contributors GitHub credits as "built by". It covers the same daily, weekly and monthly windows GitHub itself offers on the page, plus the language and spoken-language filters from the site's own UI. No GitHub account, no personal access token, and no risk of hitting GitHub's REST or GraphQL API rate limits, because this scraper never calls those APIs at all, it reads the same public page a logged-out visitor sees.

Overview

Input is a single run configuration: pick repositories or developers, a time window, and optional language filters. Output is one JSON record per trending entry, in the order GitHub's page lists them, which is GitHub's own relevance ranking and is not re-sorted by this actor. A daily run for the "all languages" trending page returns GitHub's full page (roughly 25 repositories) in a few seconds; the maxItems cap lets you pull fewer if you only need the top 5 or 10 for a digest.

Cost is pay-per-result: you are charged per trending item actually returned, not for the run itself, and not for an empty page. There is no browser rendering involved; the actor reads the page's server-rendered HTML directly, which keeps runs fast, typically well under 15 seconds, and cheap on compute compared to a headless-browser approach.

Supported inputs

• Repositories trending (kind: "repositories"): the default github.com/trending view. Supports since (daily, weekly, monthly), language (for example javascript, python, go, rust, or left empty for all languages), and spokenLanguage (an ISO code such as en, pt, ja, matching GitHub's own "trending in this spoken language" filter).
• Developers trending (kind: "developers"): the companion github.com/trending/developers view, with the same since and language filters. Returns the ranked list of developers GitHub currently features, each linked to their profile and to the repository GitHub is crediting them for on that card.
• maxItems: a cap on records returned. GitHub's own page shows roughly 25 per view; setting this higher than that simply returns everything the page has, it does not unlock more.
• debug: verbose run logs for troubleshooting, off by default.

Use cases

• Developer relations and content teams publishing a recurring "trending this week" roundup without manually visiting GitHub every time it is due.
• VCs and tech analysts tracking which repositories and languages are gaining unusually fast star velocity as an early read on developer interest, ahead of it showing up in slower-moving signals.
• Recruiters and talent-sourcing teams using the developers feed to find engineers GitHub itself is currently spotlighting, as one more sourcing channel alongside LinkedIn and job boards.
• Newsletter and Slack-bot builders piping the daily feed straight into a digest or channel post, without maintaining an in-house HTML parser against a page that changes without warning.
• Market and competitive research teams watching which libraries in a given language ecosystem are breaking out, to flag emerging dependencies before they become standard.

How it works

The actor requests the public trending page for the chosen kind, applies the since, language and spokenLanguage query parameters that GitHub's own UI uses for the same filters, and parses the resulting HTML for the repository or developer cards GitHub renders server-side. Each card's name, description, star and fork counts and today's star delta are read straight from the page markup, so the numbers match exactly what a human visiting github.com/trending would see at that same moment. There is no GitHub API call anywhere in the pipeline, so GitHub's authenticated rate limits never come into play, and there is nothing to authenticate in the first place.

Input configuration

Example 1, daily trending repositories, all languages

{
  "kind": "repositories",
  "since": "daily",
  "maxItems": 25
}

Example 2, weekly trending Python repositories

{
  "kind": "repositories",
  "since": "weekly",
  "language": "python",
  "maxItems": 25
}

Example 3, daily trending developers

{
  "kind": "developers",
  "since": "daily",
  "maxItems": 25
}

Example 4, monthly trending repositories among Portuguese-speaking projects

{
  "kind": "repositories",
  "since": "monthly",
  "spokenLanguage": "pt",
  "maxItems": 25
}

Output sample

Real output from a repositories / daily run (2026-11-06, run Zx00xhhPsFFy9LeSm):

{
  "kind": "repository",
  "author": "NVIDIA",
  "name": "auto-ontology",
  "fullName": "NVIDIA/auto-ontology",
  "description": "Auto Ontology is a Python framework for building semantic, graph-based reasoning layers over data. It enables deterministic query generation, better context retrieval, and more explainable AI for LLM and agent workflows.",
  "language": "Python",
  "languageColor": "#3572A5",
  "stars": 8076,
  "forks": 649,
  "starsToday": 2679,
  "builtBy": [],
  "url": "https://github.com/NVIDIA/auto-ontology"
}

A second real record from the same run, showing the fields populated differently depending on how GitHub's page lists the repo:

{
  "kind": "repository",
  "author": "mattpocock",
  "name": "skills",
  "fullName": "mattpocock/skills",
  "description": "Skills for Real Engineers. Straight from my .agents directory.",
  "language": "Shell",
  "languageColor": "#89e051",
  "stars": 315922,
  "forks": 26436,
  "starsToday": 2071,
  "builtBy": [],
  "url": "https://github.com/mattpocock/skills"
}

Key output fields

Field	Type	Description
kind	string	"repository" or "developer", matching the requested input kind.
author	string	Repository owner's GitHub username or organization.
name	string	Repository name.
fullName	string	owner/repo, convenient as a unique key.
description	string	Repository description as shown on the trending card.
language	string	Primary programming language GitHub detected for the repo.
languageColor	string	Hex color GitHub assigns to that language, matching github.com's own language dots.
stars	integer	Total star count at scrape time.
starsToday	integer	Stars gained within the selected since window, the metric that actually drives the ranking.
forks	integer	Total fork count.
builtBy	array	GitHub usernames credited as contributors on the trending card; can be empty when GitHub shows none.
url	string	Direct link to the repository.

For kind: "developers" runs, records instead carry username, displayName, sponsorable and url for the developer, alongside whatever repository reference GitHub's developers page links against that entry.

Pricing

Pay-per-event, $0.001 per trending item returned (item-scraped). A full daily page (around 25 items) costs about $0.025; there is no separate charge for the run itself, and an empty result page costs nothing.

What makes this richer than the competition

Two other public actors cover the same page. automation-lab/github-trending-scraper (46 users/30d) covers repositories only, with language and time-range filters, but has no developers mode and no spoken-language filter. rupom888/github-repository-scraper (23 users/30d) bundles trending into a broader GitHub repo and profile scraper at $0.0005 per result, which is cheaper per item, but trending is a side feature of a general-purpose tool rather than a dedicated feed with the full filter set GitHub's own page exposes. This actor is the only one of the three that covers both the repositories and developers trending views with the spoken-language filter in the same input schema.

Notes & limitations

This actor reads github.com/trending's public HTML, not the GitHub REST or GraphQL API, so the fields available are limited to what that page itself renders; it does not include, for example, issue counts, CI status or commit history. GitHub does not paginate trending beyond roughly 25 items per view, so a maxItems value above that returns everything the page has, not more. For the developers kind, GitHub's own page exposes less structured detail per developer than it does for repositories, since it links each developer to a single highlighted repo reference rather than a full repo card; for full repository-level detail on a developer's work, run this actor in repositories mode against that developer's own repos separately. This actor does not authenticate with GitHub and does not use or require a personal access token.

FAQ

Does this need a GitHub account or token? No. It reads the public trending page, the same one a signed-out visitor sees.

Can I get more than roughly 25 results per run? No, that is GitHub's own page limit for a single trending view; run multiple since or language combinations if broader coverage is needed.

What happens if GitHub has no trending developers for a filter combination? The dataset comes back empty and no items are charged.

Does language accept any string? It should match a language GitHub recognizes on its own filter dropdown, such as javascript, python, typescript, go, rust or java. An unrecognized value falls back to "all languages".

How fresh is the data? As fresh as github.com/trending itself, which GitHub updates multiple times per day.

Can I filter by more than one language in a single run? No, one language value per run; run it again for a second language.

For AI Agents & LLM Apps

Call this actor from any MCP-compatible client via https://mcp.apify.com, or synchronously from any HTTP client:

curl "https://api.apify.com/v2/acts/gio21~github-trending-scraper/run-sync-get-dataset-items?token=$APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"kind": "repositories", "since": "daily", "maxItems": 10}'

This returns the dataset items directly in the response, which is ideal for handing an LLM agent a same-call snapshot of what is trending on GitHub right now, without a separate poll-for-completion step.

Disclaimer

This Actor is an independent tool and is not affiliated with, endorsed by, or sponsored by GitHub, Inc. or any of its subsidiaries. All trademarks mentioned are the property of their respective owners.

SEO Keywords

GitHub trending scraper, GitHub trending API, scrape GitHub trending repositories, GitHub trending developers scraper, GitHub star tracker, open source trend tracking tool.