JADX plugin and MCP server enabling live LLM reverse engineering for Android APKs.
Bridge JADX decompiler with Claude and other LLMs via Model Context Protocol.
JADX-AI-MCP is an open-source suite combining a JADX plugin and a Python MCP server. It connects decompiled Android app contexts directly to LLM clients like Claude for real-time security analysis and reverse engineering.
The tool uses the Model Context Protocol to let LLMs fetch classes, methods, and resources from JADX, search codebases, read AndroidManifest files, and execute refactoring commands. It also integrates with the JADX debugger to pull stack frames, threads, and variables.
It suits security professionals, malware analysts, and reverse engineers working with Android applications. It does not suit developers looking for standard IDE assistance without binary decompilation.
Connects JADX directly to LLM clients to analyze decompiled Java and Smali code in real time.
Provides over two dozen tools to fetch classes, search methods, read manifests, and query resource files.
Extracts stack frames, variables, and thread insights directly from the JADX debugger state.
Allows LLMs to rename classes, methods, fields, and variables or add persistent project comments.
Free under the Apache-2.0 license. The only cost is running the software yourself and any API fees associated with your chosen LLM client.
873 open remote AI-training roles, updated November 15, 2026; most pay $20–$170 an hour. Zekai lists them free; the application goes to the employer’s platform.
You can use any LLM client that supports the Model Context Protocol, such as Claude.
No, modifications like renames and comments are saved within the JADX project context.
You need Java 11+ for the JADX plugin and Python 3.10+ for the MCP server.
See Zekai first in Google