Why iCR


Generates accurate fixes not suggestions


Generates low false positive rate. ICR < 3% vs.others >50%


Increase code quality automatically at lower cost


Improves developer productivity by 11%


Focuses resources on features instead of bug fixing

What Drives iCR

iCR reduces to practice cutting-edge research done in three prominent disciplines.

Deep Static Analysis

Deep pointer analysis creates accurate knowledge vital to code repairing

Machine Intelligence

Machines learn from your code to generate fixes as if done by a human developer

Code Refactoring

Behavior enhancing refactoring transformations fix problems while preserving original behavior

How does it work


  • Runs on AWS EC2 instances
  • Accesses source code securely using GitHub or GitLab
  • Fixes turned into pull requests
  • Securely runs on your on-premise servers using Docker
  • Access source code from private repositories
  • Easy to use interface


What does it fix

57 Fixers that cover

  • API Usage Issues
  • Arithmetic Issues
  • Bad Control Flow
  • Concurrency Issues
  • Improper Access Control
  • Improper Method Call
  • Null Pointer Issues
  • Object Visibility
  • Security Misconfiguration Issues
  • Sensitive Data Exposure
  • Cross Site Request Forgery Issues
  • Weak Cryptography Issues