What is CodeAi?

CodeAI is an Intelligent software debugging system that automatically finds and fixes security vulnerabilities and other types of bugs in computer source code.

How does CodeAi find and fix bugs?

CodeAI takes a unique approach to finding and fixing bugs by learning from previous real-world bugs and fixes in large amounts of software. We collect bug-fix examples in open source C, C++ and Java repositories on GitHub, GitLab, etc. From these repositories, it analyzes the prefix and postfix source files for differences that are indicative of security vulnerability bug and fix patterns. The patterns are encoded as feature vectors that are fed into a machine learning model, which learns to correlate new bugs with likely fix actions.


How does CodeAI validate it's fixes?

CodeAI validates the correctness of each fix that it generates using a combination of static analysis, automated test suite and state of the art AI (deep learning). This verifies that the modified program builds, that the detected defect was removed from the static analysis results, and that no new defects get introduced. Additional validation configurations can be made to verify the correctness of the fix, however, the final determination of fix correctness still remains in the hands of users who know their code.

How capable is CodeAi now?

There are currently ~30 CWE categories of issues that every software project is dealing with. CodeAi is capable of fixing all of them in a matter of seconds.

What languages does it support?

CodeAi is designed to support top 10 programming languages, including C and C++ with Java and C#.

If you usue GitHub, GitLab etc, all yo need to do is to login with your GitHub login.

How easy it is to use?