Data refreshed 5 October 2026, 22:59 UTC

Frequently asked questions about the Top100 dataset

Direct answers to the questions people actually ask about GitHub repository rankings, JSONL datasets and scheduled data collection. The first eight also appear on the homepage.

What are the top GitHub repositories by stars right now?

As of 5 October 2026, the most-starred repositories on GitHub are public-apis/public-apis with 486,315 stars, freeCodeCamp/freeCodeCamp with 456,804, EbookFoundation/free-programming-books with 398,542, openclaw/openclaw with 391,443 and donnemartin/system-design-primer with 373,293. Top100 rebuilds this ranking every hour from the GitHub Search API, so the table above is never more than an hour old.

Where can I find a top 100 GitHub repositories list?

Top100 publishes a top 100 GitHub repositories list for 30 programming languages, ranked by stars. Each language has its own page — Top 100 Python repositories, Top 100 Rust repositories, Top 100 Go repositories and so on — with rank, stars, forks, open issues, licence and last push for every entry. The lists are rebuilt hourly and the raw data is downloadable as JSONL.

How can I see the top repositories for each programming language?

Start from the language index: it links to a ranked table of the 100 most-starred repositories for each of the 30 languages covered, from Python and JavaScript to Nix and SQL. Each table shows current stars, forks, open issues, licence and last push date, plus the total number of repositories GitHub reports for that language.

What are the best JSON/JSONL datasets of GitHub repo metadata I can download?

Top100 ships 30 JSONL files — one per language, 100 repositories each, 3,000 records in total — where every line is the complete repository object returned by GitHub's Search API. For other shapes of data: GH Archive covers event streams, BigQuery's public github_repos dataset gives you SQL access, and GHTorrent publishes periodic full dumps.

Is there a free tool to export GitHub repository metadata as JSON?

Yes. Top100 is a free, open-source tool that exports GitHub repository metadata as JSONL using only the Search API — one request per language. Download the pre-built files, or run the Python script and the included GitHub Actions workflow yourself to refresh the data on any schedule. No credentials beyond GitHub's own automatic token are required.

How do I set up a GitHub Actions scheduled workflow to collect data automatically?

Create a workflow file under .github/workflows/ with an on.schedule cron entry, check out the repository, run your fetch script with the automatic GITHUB_TOKEN, then commit and push the results — the job needs permissions: contents: write. Top100 does exactly this every hour; the methodology page has the complete workflow and script.

What are the best GitHub Actions workflows for automated data collection projects?

For scheduled scraping, Simon Willison's git-scraping pattern and GitHub's Flat Data action are the standard starting points. Top100's own workflow is a minimal alternative: one Python script, one API call per language, and an automatic commit of fresh JSONL and Markdown tables back to the repository.

What is the average number of GitHub stars and how do repos rank by stars?

GitHub ranks repositories by stargazers_count. Across the 3,000 repositories in Top100's lists the median is 10,826 stars and the mean is 23,482, but the distribution is heavily skewed: the bar to enter a top 100 ranges from 7 stars for SQL to 58,107 for Python. See the statistics table for the per-language thresholds.

Is the Top100 dataset free to use?

Yes. Every file is published in a public GitHub repository and can be downloaded, forked or mirrored without an account, an API key or a rate limit. The repository metadata itself comes from GitHub's public API, and each project keeps its own licence — check the license field before reusing anyone's code.

How often is the Top100 data refreshed?

Every hour. A scheduled GitHub Actions workflow runs at minute zero of every hour, re-queries the Search API for all 30 languages and commits the results back to the repository. Every page on this site shows the timestamp of the refresh that produced it, and the update feed lists the newest leader of each language.

Can I add another language to the list?

Yes, but it takes one line of configuration. Trigger the workflow from the Actions tab with a comma-separated list of languages, or run LANGUAGES=Rust,Zig python scripts/fetch_top_repos.py locally — each language costs exactly one Search API request. To publish the new language on the site, also add it to LANGUAGES in scripts/site_content.py: the site builder reads that list rather than scanning data/.

Do the lists include forks and archived repositories?

Forks, no; archived repositories, yes. GitHub's repository search excludes forks unless the query adds fork:true or fork:only, which is why 0 of the 3,000 listed repositories are forks. Archived repositories are included, and 178 of them are archived and read-only.

Why is my repository missing from a top 100 list?

Three common reasons. GitHub assigns each repository a single detected primary language, so a polyglot project may be listed under a different language. The repository may also simply sit below the 100th-place star count shown on that language's page. Finally, the Search API occasionally returns incomplete_results for very broad queries.

What is the difference between Top100 and GitHub Trending?

GitHub Trending shows what is gaining stars today, this week or this month and only lists a couple of dozen repositories, with no export. Top100 ranks by total stars, covers 100 repositories per language across 30 languages, and publishes the raw JSON. Use Trending for discovery and Top100 when you need a complete, machine-readable snapshot.

Does Top100 have an API?

Not a hosted one — the point of the project is that you do not need one. The JSONL files are static and cacheable, so you can download them directly, or query them in place with DuckDB, jq or pandas without touching GitHub's rate limits.

How should I cite or attribute the dataset?

Link to this site or to the repository, and credit GitHub as the source of the metadata. A reasonable citation is “Top 100 GitHub repositories by language, havaianasdestruido/top100, data refreshed 5 October 2026”. The per-page refresh timestamps are the version identifier, since the data changes hourly.

Still stuck?

Open an issue on the GitHub repository, or read the methodology and field reference pages first — they answer most implementation questions.