Gain Expert Insights into your JavaScript and TypeScript Applications with a Knowledge-Driven Static Code Analyzer
Just read the automatically generated explanations of what the code means and does, the warnings of errors and security vulnerabilities, and the possible fixes
NALAdocs are knowledge bases, a new form of programming documentation written by domain experts to bring their knowledge into your code editor.
These knowledge bases include the definitions of domain concepts, programming errors arising from language and API missuses, security issues, and possible fixes.
The NALAdocs Reader applies this information to your code to explain its meaning and behaviour and alert you of any defects and vulnerabilities.
Unlike conventional rule-based and pattern-matching static code analysis tools, NALAdocs Reader is a knowledge-driven expert system. It builds a novel concept model of the source code and employs an inference engine to reason about your code.
Think of it as virtual pair-programming with those domain experts - to save you time, quickly bring you up-to-speed with unfamiliar code, and help you produce better and safer software.
Cutting-edge Static Code Analysis
Privacy
Runs locally without
using external services
No Hallucinations
Deterministic knowledge-
based analysis
Robust
Works even with code
that does not compile
Fast
Analyses source files
in near real-time
Code reviews are an essential part of software development to detect and correct coding mistakes, security vulnerabilities and other quality issues before they reach production.
Code reviews are also necessary for AI-generated code. AI has changed how fast you can produce functional code in minutes instead of hours, but this results in code overload because the speed of manually reviewing and understanding that code has not changed. This has become the new bottleneck.
As a first line of defence against codding errors in human written and AI-generated code developers employ static code analysis tools in their code editors to detect and correct those errors. Unique among static code analysis tools, NALAdocs are also code explainers that address that bottleneck by helping you to understand and come up to speed with that code in a lot less time.
Understanding AI-generated
Code is the new Bottleneck
Trusting code without fully grasping how it works can introduce hidden bugs, security flaws, and maintenance challenges, but manual reviews can be difficult if you don't fully understand the code.
Each source code construct can have many different meanings and codding issues which can’t be inferred from just the syntax. The NALAdocs Reader surfaces all that knowledge for you. It builds a layered cross-domain model of the code to explain everything from the syntax and operational semantics of each language construct, the application semantics of library calls, through to the coding errors and security flaws in each domain of discourse.
AI-generated code that calls external APIs or libraries is more challenging to interpret than human-written code so demands even deeper understanding. The NALAdocs Reader addresses this need by employing advance static analysis techniques such as cross-file data flow tracking and cross-file taint analysis.
It’s a novel form of static analysis that enables you to understand all aspects of your code without having to manually hunt for that information across multiple documentation sources.
Layers of explanations
that can't be inferred
The NALAdocs Reader tracks calls to web services, the query parameters you set and default values to explain what those services will do.
That information filters down through the source code model, mapping to the libraries which call out to web services, variables used to specify query parameters, and the callback functions that handle returned data and possible errors.
The result is precise explanations about how your code interacts with the web, so you can understand the cloud like never before.
Understand how your application
interacts with web services
A Community Platform
NALAdocs represent a paradigm shift in static code analysis tools by the way we model source code based on programming knowledge. We leverage this model to produce Natural Language Annotations (NALA) of the code to explain its behavior including errors and vulnerabilities, but there are many other possible uses of this source code model.
Community involvement is key to developing the full potential of this new technology. We plan to expose our APIs, and later we also aim to open source our software. If you would like to get involved, sign up for our newsletter and we’ll keep you posted of developments.