A sample of real source repositories from the data lake, packed into these pages in full. Open any project and read its entire tree with syntax highlighting, not a snippet, the whole thing. The selection leans toward C/C++ and Java, the code most researchers come here for.
Across 244,921 distinct source repos, 98% are not on GH Archive; only about 2% show up there, and those are re-uploads of already-public projects. Everything shown here is from that novel set, so it isn't code you could pull off GitHub.
What "embedded" means: each page carries a spread of complete repos for that language, small, medium, and large, packed straight into the HTML, so they open instantly with no network calls. It's a representative selection you can read end to end, not the entire on-disk sample. Every repo here is complete: each file you can open carries its full body, top to bottom, no partial repos. Non-source artifacts (config like xml/json, minified vendored libraries, binaries and image assets) are not shown; the authored source itself is whole. The full repos, artifacts and all, are in the S3 download.