CodeAsk  by woniu9524

LLM-based code reader for legacy, obfuscated, or undocumented codebases

created 6 months ago
906 stars

Top 41.0% on sourcepulse

GitHubView on GitHub
Project Summary

CodeAsk is an LLM-powered desktop application designed to help developers understand and maintain complex codebases. It addresses common pain points like poorly documented or legacy code by providing AI-driven analysis, security vulnerability detection, and code quality assessments. The target audience includes individual developers and teams struggling with code comprehension, aiming to boost productivity and reduce maintenance overhead.

How It Works

CodeAsk leverages Electron for its desktop application framework, integrating React for the UI and Monaco Editor for code viewing. It utilizes Large Language Models (LLMs) to perform code analysis, with users able to configure various models and custom prompts via a plugin system. The application supports both single-file and global code analysis, generating reports that can be viewed within the application or shared via a .codeaskdata file. An IntelliJ plugin is also available for IDE integration.

Quick Start & Requirements

  • Install dependencies: npm install --legacy-peer-deps
  • Start application: npm run start
  • Requires Node.js >= 16 and React ^19.0.0.
  • LLM API keys and compatible base URLs are necessary for model configuration.
  • Official Demo Video: https://bilibili.com/video/BV1eQNzeuEEb/

Highlighted Details

  • AI-powered code analysis for understanding complex code.
  • Supports custom LLM configurations and prompt engineering.
  • Includes a read-only code preview with theme and font customization.
  • Offers an IntelliJ plugin for in-IDE code analysis.

Maintenance & Community

  • The project is actively developed, with plans for VS Code plugin integration.
  • Users are encouraged to submit issues for bugs and share prompt templates.
  • Discussion and prompt sharing can occur via GitHub issues.

Licensing & Compatibility

  • Licensed under GNU (GPL).
  • The GPL license may impose copyleft restrictions on derivative works, potentially requiring shared source code if linked or distributed.

Limitations & Caveats

The project is primarily developed in Chinese, with English and Japanese translations provided. Prompt templates are largely generated by DeepSeek and may require user testing and adjustment based on specific models and code. The IntelliJ plugin is not yet available on the official marketplace and requires manual installation.

Health Check
Last commit

3 months ago

Responsiveness

1 day

Pull Requests (30d)
0
Issues (30d)
0
Star History
44 stars in the last 90 days

Explore Similar Projects

Feedback? Help us improve.