Today we are excited to announce the general availability of Code Sa for C/C++ open source projects.

Starting today, anyone can use Code Sa to find all the users of a C/C++ software library on GitHub. Just head to beta.code-sa.ai and try it out for yourself. After you log in with your GitHub credentials, Code Sa automatically lists every publicly visible C/C++ project you have access to. From there, you can select a repository and start gathering analytics on test coverage and on how your library is used across all of GitHub.

To go along with the launch, we've also just released our white paper, which explains what Code Sa is and what it can do today. It's our first step towards fully autonomous code maintenance.

———————

TL;DR

Your application depends on code that no one fully tests. A single bug in a widely used library can ripple out to thousands of application.

Code Sa closes that gap in two steps

Visibility

With a single click, Code Sa shows exactly which library functions are used and how well each is tested, turning "usage vs. coverage" into one clear picture of where your risk really is. This serves two audiences:

If you build applications

Code Sa identifies which functions you depend on from a library and reports each one's current test coverage, based on the library's own tests, giving you immediate visibility into your exposure to potential bugs.

If you maintain libraries

You already know your overall coverage numbers, but which functions in your API are most popular? Code Sa shows who depends on what, and how real-world usage lines up against your testing strategy.

Action

Code Sa generates tests grounded in how applications actually use a library, rather than the contrived, surface-level tests of traditional tools. The whole pipeline runs autonomously, from finding coverage gaps to handing you a single-click pull request.

The results back it up: benchmarked against Claude Code on the same untested functions, Code Sa added more line coverage across all three libraries tested, including over 1,000 extra lines on FFmpeg alone.

Our test generation component is currently available through limited trials only. If you're interested, please reach out —we'd love to hear from you.

Try it now at beta.code-sa.ai and read the full white paper to see where we're headed with autonomous code maintenance.