cncsnet1/jadx-gui-ai

What it solves

This project enhances the popular JADX-GUI tool with AI capabilities to simplify Android application reverse engineering. It helps users analyze APKs by automatically identifying technical stacks, mapping API call paths, and providing evidence-based insights, reducing the manual effort required to navigate complex codebases.

How it works

It integrates an AI agent into the JADX-GUI interface. When an APK is loaded, the tool generates a local report and identifies potential investigation routes. Users can then configure AI connections (supporting various LLM providers) to perform deep analysis. The agent reads relevant source code snippets, maps call chains, and organizes findings into a structured investigation workbench where users can verify evidence, add notes, and track progress.

Who it’s for

It is designed for security researchers, developers, and reverse engineers who need to analyze Android applications but want to streamline the process of finding entry points, understanding network communication, and auditing sensitive code sections.

Highlights

• Goal-Driven Analysis: AI-powered task workbench that plans investigation steps and retrieves evidence from local source code.
• Integrated Workflow: Automatically maps API calls, network requests, and business logic paths without requiring external plugins.
• Evidence-Based: Provides traceable links between AI conclusions and the specific source code lines used as evidence.
• Flexible Connectivity: Supports multiple AI model configurations, allowing users to switch between local and cloud-based models for different tasks.
• Burp Suite Integration: Includes an MCP server to link static analysis with dynamic traffic captured in Burp Suite.