code-archive

Software Heritage

A universal archive of source code, the 'Library of Alexandria of code', preserving over 20 billion unique source files from more than 350 million projects. You can search for projects, look up files by hash (sha1/sha256/git), and resolve permanent SWHID identifiers, on the website or through its keyless REST API. The place to go when a repository has vanished from its original host, or when you need to identify a file by its hash.

Software / packages API

Search this archive

Why it’s useful & how it works

Software Heritage is the closest thing to a universal safety net for source code: it continuously crawls public repositories from GitHub, GitLab, package managers like PyPI and npm, and dozens of other forges, preserving every unique file and commit with a permanent SWHID identifier. A researcher reaches for it when a repository has disappeared from its original host, when they need to pinpoint an exact file by its cryptographic hash, or when they want a stable, citable reference to a specific version of code. You can search by project name or URL to find archived origins, look up a file directly by its SHA-1, SHA-256, or git hash, or resolve an SWHID to jump straight to a preserved snapshot.

What’s inside

Over 20 billion unique source files and 350 million software projects preserved, drawn from forges, package repositories, and FOSS distributions worldwide. The archive crossed 2 petabytes of stored data in 2025.

API access

https://archive.softwareheritage.org/api/1/ , /content/sha1:<h>/, /content/sha256:<h>/, /origin/search/<text>/, /resolve/<SWHID>/

What we measured

Our own probes, not the archive’s own claims. Re-run periodically; every reading below is dated.

Reachability

Direct request: Responded HTTP 200 338 ms
Through a datacenter proxy: Responded HTTP 200 767 ms
API, direct: Responded HTTP 200 209 ms
API, through a proxy: Responded HTTP 200 770 ms

Did a real search return anything?

Returned results · 10 results for the sample query · median 1 s · stable across runs

Sample query: https://github.com/DzRepo/Python-Insights

Benchmark coverage

Found 9 of 10 test items (90%) in the Software / packages cluster. Median response 532 ms.

By how well known the item is: 5/5 well-known, 4/4 mid, 0/1 obscure .

One item it found: https://github.com/mehul-clou/sending-http-requests-starter

Found 0 of 2 test items (0%) in the File by hash cluster. Median response 592 ms.

By how well known the item is: 0/2 mid .

Reachability measured 2026-08-22. Search checked 2026-08-22. Benchmark run 2026-08-22.