We fix code
to make software safer
We have created an entirely new class of technology that works hand in hand with engineers to accelerate the delivery of trusted software at scale
Detects Bugs that are Invisible to Other Tools
Using a new class of algorithms that allow us to model and evaluate billions of execution paths at unprecedented depth, our intelligent code repair technology eliminates software flaws that are too complex for humans and traditional static testing tools to find. Using an engine that describes the fundamental structures of popular software frameworks such as Spring,
Spring Boot, and JEE, our products are able to understand the code paths of these frameworks with a depth and accuracy far beyond the capabilities of traditional static analysis tools that rely on general heuristics rather than deep individual analysis. By leveraging this information for deep learning, our tools are able to find classes of bugs that other tools simply cannot see.


Click HERE to see how iCR detected bugs that others missed

Use Case #1
How iCR detected bugs that others missed

In the recent past companies have suffered from several high impact vulnerabilities. From the Heartbleed bug in OpenSSL, to the LDAP injection vulnerability in Log4J, through to the sensitive data leak in the ctx library of the PyPI repository.

In each case our intelligent code repair technology was uniquely equipped to find the bugs that caused these vulnerabilities.

This is backed up by evidence from the SAMATE benchmark

maintained by the National Institute of Standards and Technology (NIST), where our technology was consistently able to find around two times more bugs than traditional static analysis tools.

Delivers a Dramatic Reduction in False Positives
Using a new set of high precision models that understand the full context of your code, our intelligent code repair technology is able to ensure that discovered bugs are true errors within your specific code and not false positives caused by statistics-based assumptions. By combining deep analysis, machine learning, and natural language processing techniques, our products are able to process the execution paths of your software with full understanding
of its context. This enables our tools to discover vulnerabilities in your software with an accuracy and correctness far beyond that achievable in traditional static analysis tools that rely only on statistical models and heuristics. Ensuring that developers do not have to waste significant time and effort dealing with hundreds – or even thousands – of inaccurate bug warnings.


Click HERE to see how iCR detected bugs with dramatically low false positives

Use Case #2
How iCR delivered a dramatic reduction in false positives
Many companies have rolled out static analysis tools as part of their development process but remain unsatisfied by the numbers of false positives generated by traditional static analysis.

The breakthrough technology used by our intelligent code
repair (iCR) technology was explicitly built to overcome this challenge and transform accuracy and productivity.

The advantage that we have achieved is again backed up byevidence from the SAMATE benchmark, with our deep analysis technology
generating less than 5% false warnings compared to over 70% generated by the statistical models applied by traditional static analysis tools.
Provides AI-based Code Correction for Developers
Using a range of error correction modules that encode expert knowledge, our intelligent code repair technology is able to automatically correct bugs in your code by using machine learning to eliminate errors with a coding style that mimics your own. Making our intelligent code repair the only tool that can automatically synthesize code in order to fix problems on your behalf. Using deep source code analysis in combination with machine
learning, our intelligent code repair technology is able to create human-level fixes in a constantly growing range of frameworks and languages. Individual fix modules can focus on correcting specific bugs, collections of files, metadata, or even configuration data. Proposed fixes are then presented to developers so that they can either choose to have code corrections applied automatically or use the presented context to apply their own manual fix.


Click HERE to see how iCR synthesized fixes automatically

Use Case #3
The efficiency of iCR compared to other tools
In the vast majority of cases, even the most pernicious and dangerous bugs can have a relatively simple fix.

Our goal has not only been to discover these bugs with unprecedented precision but also to augment the expert
capabilities of developers by enabling them to save time with automated code refactoring and fixes.

During empirical testing our intelligent code repair technology was consistently able to synthesize fixes for
over half of the bugs found, with a rapid improvement path towards 75%.

For bugs that require developer input, our technology provides all related context about the bug to help them close it faster.

bug discovery
and repair

Our intelligent code repair
catches and fixes a wide variety of different bug types, whether the risks are related to security, reliability, compliance, or logic.

Security Bugs

We protect your application’s integrity and reputation by eliminating security problems such as injection attacks, weak authentication, or weak cryptography.

Compliance Bugs

We protect your developers from technical debt by ensuring our automated code fixes are fully compliant with the coding practices you use to ensure maintainability.

Reliability Bugs

We protect your users from poor experiences by eliminating reliability problems that can bring down your system such as null pointers, race conditions, or memory leaks.

Logical Bugs

We protect your development velocity by using deep analysis and integrated tool support to help developers discover and eliminate the most pernicious logical bugs.