Code-to-Spec
  • 20 Dec 2024
  • 1 Minute to read
  • Dark
    Light
  • PDF

Code-to-Spec

  • Dark
    Light
  • PDF

Article summary

This functionality is available on specific configurations. If it is not available in your system, contact our Support team.

Overview

The Code-to-Spec functionality helps you modernize your older applications and codebases by importing an existing GitHub repository and converting it to a PRD. Code-to-Spec supports a wide range of file types, including popular programming languages, configuration files, and markup languages.

The functionality is especially useful for understanding and updating codebases that are outdated, difficult to interpret, or written in older languages like COBOL. It helps teams efficiently analyze and modernize their code, streamlining upgrades, language conversions, and maintenance.

After importing the repository, the platform provides detailed insights into the code's structure, functionality, and key components.

During the process, the platform performs the following tasks:

  • Clones the repository remotely.

  • Chunks the codebase into manageable pieces.

  • Interprets the structure and purpose of the files and folders.

  • Summarizes each file (in parallel).

  • Compiles the summaries into a final summary.


The platform then uses the final summary to create a PRD, which can be used to guide further development, upgrades, and maintenance.

For more information, refer to Using Code-to-Spec.


Supported Languages

  • .Net Core:
    • Visual Basic
    • Visual C++
    • C#
  • COBOL
  • Java
  • Javascript
  • PHP
  • Python
  • Typescript

Additional Support

Support for other languages and frameworks is being added frequently. Talk to your Crowdbotics Account Executive if your framework isn’t shown in the list above.


Best Practices

  • Smaller repositories generally perform better during analysis, making it easier for your teams to optimize their workflows and interpret results.

  • To ensure the best results, Code-to-Spec adheres to specific limitations, such as analyzing directories up to 5 levels deep and processing individual directories under 0.6 MB in size.



Was this article helpful?