JADX-AI-MCP

Project

Author: zinja-coder
Repository: github.com/zinja-coder/jadx-ai-mcp
License: Apache-2.0
Category: Security & Reverse Engineering

Tech

Runtime / language: javaJava and Python
Transports: stdio
Remote available: No
Authentication: Not specified

Activity

GitHub stars: 2,909
Last commit: Sep 23, 2026
Data as of: Oct 16, 2026

Overview

JADX-AI-MCP links assistant conversations to an Android application opened in the JADX decompiler. It is a two-part integration: a Java plugin inside JADX and a companion Python MCP server. The assistant calls the server, which forwards requests to the plugin and returns information from the live GUI context. This arrangement supports reverse engineering and security review without requiring every relevant class or resource to be copied manually into a conversation.

The available tools expose selected text, class and method source, fields, smali, application classes, and Android manifest content. Resource operations retrieve strings and other packaged files. Search and cross-reference tools help trace behavior across classes, methods, and fields, with pagination documented for several operations. Renaming tools cover packages, classes, methods, fields, and local variables. Debugger access adds stack frames, threads, and variable state, allowing the assistant to reason about runtime observations alongside decompiled code.

Setup requires both the JADX plugin and its companion server, not just the plugin artifact. The README documents release downloads, a plugin installation command, uv-based server execution, and an executable installation option. It lists Java 11 or newer and Python 3.10 or newer. Stdio is the basic client connection, with HTTP mode available separately. Client-facing listening settings and the target plugin address are independent, which matters when the server and JADX run on different machines.

Install & configure

Requirements

Java 11 or newer
Python 3.10 or newer
JADX with the JADX-AI-MCP plugin installed
Companion JADX MCP Server
uv

Documented launch / install command

uv --directory path/to/jadx-mcp-server run jadx_mcp_server.py

This follows the README's Cherry Studio configuration. Replace the directory with the companion server checkout. The plugin and server are separate components; uv venv alone does not start the MCP server.

Claude Desktop

Cursor

VS Code

Windsurf

Claude Code

Config file · macOS: ~/Library/Application Support/Claude/claude_desktop_config.json

Windows: %APPDATA%\Claude\claude_desktop_config.json

Use Settings → Developer → Edit Config when available. Do not assume a Linux configuration path; check the client documentation for your platform. Remote servers use the app’s connector flow when supported, not the local stdio file shown here. Merge this example with existing settings; do not overwrite unrelated servers.

Claude Desktop configuration

{
  "mcpServers": {
    "jadx-ai-mcp": {
      "command": "uv",
      "args": [
        "--directory",
        "path/to/jadx-mcp-server",
        "run",
        "jadx_mcp_server.py"
      ],
      "env": {}
    }
  }
}

Read the Claude Desktop setup guide

Config file · Project: .cursor/mcp.json

User: ~/.cursor/mcp.json

The project path is relative to the workspace root. The tilde denotes your user home directory. Use Cursor’s MCP settings to confirm which configuration and scope your installed version is reading. Merge this example with existing settings; do not overwrite unrelated servers.

Cursor configuration

{
  "mcpServers": {
    "jadx-ai-mcp": {
      "command": "uv",
      "args": [
        "--directory",
        "path/to/jadx-mcp-server",
        "run",
        "jadx_mcp_server.py"
      ],
      "env": {}
    }
  }
}

Read the Cursor setup guide

Config file · Workspace: .vscode/mcp.json

For user scope, use the “MCP: Open User Configuration” command if available. Do not guess a user configuration path across operating systems, profiles, containers, or remote workspaces; check the client documentation. Merge this example with existing settings; do not overwrite unrelated servers.

VS Code configuration

{
  "servers": {
    "jadx-ai-mcp": {
      "type": "stdio",
      "command": "uv",
      "args": [
        "--directory",
        "path/to/jadx-mcp-server",
        "run",
        "jadx_mcp_server.py"
      ],
      "env": {}
    }
  }
}

Read the VS Code setup guide

Config file · macOS/Linux: ~/.codeium/windsurf/mcp_config.json

Windows: %USERPROFILE%\.codeium\windsurf\mcp_config.json

Open the MCP configuration through Windsurf’s settings when possible to confirm the active file. Paths can differ in managed or future releases; check the client documentation if this location is not present. Merge this example with existing settings; do not overwrite unrelated servers.

Windsurf configuration

{
  "mcpServers": {
    "jadx-ai-mcp": {
      "command": "uv",
      "args": [
        "--directory",
        "path/to/jadx-mcp-server",
        "run",
        "jadx_mcp_server.py"
      ],
      "env": {}
    }
  }
}

Read the Windsurf setup guide

Config file · Project scope: .mcp.json at the repository root

The project file uses mcpServers. For user scope, use claude mcp add --scope user rather than assuming .mcp.json in your home directory is the user configuration. Check the client documentation for managed settings and storage details. Merge this example with existing settings; do not overwrite unrelated servers.

Claude Code configuration

{
  "mcpServers": {
    "jadx-ai-mcp": {
      "type": "stdio",
      "command": "uv",
      "args": [
        "--directory",
        "path/to/jadx-mcp-server",
        "run",
        "jadx_mcp_server.py"
      ],
      "env": {}
    }
  }
}

Read the Claude Code setup guide

Examples are translated from the published configuration. Supply your own credentials and verify requirements in the upstream README. Remote connectors are shown only where a supported configuration format is known.

Tools

Source: repository README. Tool descriptions are a documented snapshot, not a live capability check.

Tool	What it does
fetch_current_class	Read the selected class name and source from JADX.
get_selected_text	Retrieve text highlighted in the decompiler.
get_all_classes	Enumerate classes in the opened project.
get_class_source	Load decompiled source for a named class.
get_method_by_name	Read a method's decompiled source.
search_method_by_name	Locate matching methods across classes.
search_classes_by_keyword	Search class source for a keyword with paginated results.
get_methods_of_class	Enumerate methods belonging to a class.
get_fields_of_class	Enumerate a class's fields.
get_smali_of_class	Retrieve a class's smali representation.
get_main_activity_class	Resolve the main activity declared in the Android manifest.
get_main_application_classes_code	Read source for application classes selected through the manifest package.
get_main_application_classes_names	List application class names associated with the manifest package.
get_android_manifest	Retrieve the application's AndroidManifest.xml.
get_manifest_component	Read one manifest component rather than the full document.
get_strings	Load the packaged strings.xml resource.
get_all_resource_file_names	Enumerate packaged resource filenames.
get_resource_file	Read a selected resource file.
rename_class	Change a class's displayed name.
rename_method	Assign a clearer name to a method.
rename_field	Change a field identifier.
rename_package	Rename an application package.
rename_variable	Rename a local variable within a method.
debug_get_stack_frames	Inspect stack frames from the JADX debugger.
debug_get_threads	Retrieve thread information from the debugger.
debug_get_variables	Read debugger variable values.
xrefs_to_class	Find class and method references targeting a class.
xrefs_to_method	Find method references, including override-related methods.
xrefs_to_field	Locate methods that access a field.

Names, permissions, and defaults can change. Check the current upstream documentation before granting access.

Use cases

Review decompiled Android code for insecure API use or embedded secrets.
Follow references to understand an APK's behavior.
Rename obfuscated symbols during an investigation.
Inspect manifest entries, resources, and debugger state together.

Strengths

Connects directly to the current JADX context.
Combines source, resources, cross-references, and debugger information.
Supports both stdio and HTTP client connections.

Limitations

Requires a plugin plus a separate Python server.
Remote setups must distinguish the MCP listener from the JADX plugin connection.

Frequently asked questions

No. The integration also uses a standalone Python MCP server that communicates with the plugin.

Yes. Tools expose strings.xml, resource filenames, individual resources, and manifest content.

Yes. The documented tools retrieve stack frames, threads, and variables from JADX's debugger.

No. The MCP listener has its own host and port, while separate JADX host and port flags identify the plugin.

Independent listing, not an affiliation or security review. Grant the least access needed and verify the upstream instructions before running code.

Server facts

Project

Tech

Activity

Tags

Overview

Install & configure

Requirements

Tools

Use cases

Strengths

Limitations

Frequently asked questions

Related MCP servers

IDA Pro MCP

ghidraMCP

MCP Server for WinDbg Crash Analysis

apktool-mcp-server

Server facts

Project

Tech

Activity

Tags

Overview

Install & configure

Requirements

Tools

Use cases

Strengths

Limitations

Frequently asked questions

Is installing the JADX plugin enough?

Can it inspect APK resources?

Can an assistant see debugger information?

Are HTTP and plugin ports the same setting?

Related MCP servers

IDA Pro MCP

ghidraMCP

MCP Server for WinDbg Crash Analysis

apktool-mcp-server