exa-mcp-server
Exa's MCP server for web search and crawling — give your AI agent a live, high-quality web index.
jadx-ai-mcpは、zinja-coderが公開しているオープンソースのAIエージェントプロジェクトです(GitHub: zinja-coder/jadx-ai-mcp)。GitHubスターは2,869、フォークは264で、Apache-2.0ライセンスの下で公開されています。GitHubでの最終更新は2026年9月23日です。
JADX plugin + MCP server that lets LLMs analyze Android APKs live — decompile, review, find vulnerabilities
Have an APK you are authorized to analyze and a goal: understand its code, audit it, or hunt for vulnerabilities.
Download jadx-ai-mcp-<version>.jar and jadx-mcp-server-<version>.zip from GitHub Releases and unzip them.
Install the plugin into JADX, set up the server environment with uv, and add the server to your MCP client (e.g. Claude Desktop) config.
Open the APK in JADX, then ask your LLM to fetch the current class or search for a suspicious API and explain it.
The LLM works against the APK currently open in JADX, not stale exported files.
Tools include fetch_current_class(), get_selected_text() and more for reading classes, methods and selected code (full list in the README/docs).
Designed for the decompile → context-aware code review → AI recommendation loop for security analysis.
A JADX GUI plugin (Java) talks over HTTP to a standalone Python MCP server (jadx_mcp_server.py, using fastmcp/httpx).
Sits alongside the author's APKTool-MCP-Server and ZIN-MCP-Client projects.
Full documentation is published at jadx-ai-mcp.readthedocs.io.
Clone the plugin repo (or download the release jar + server zip from Releases).
git clone https://github.com/zinja-coder/jadx-ai-mcp.git
cd jadx-ai-mcpUnzip the server, enter its directory and install its dependencies with uv as the README shows.
cd jadx-mcp-server
uv pip install httpx fastmcpPoint your client at the server via uv run, restart the client, and open an APK in JADX.
{
"mcpServers": {
"jadx-mcp-server": {
"command": "/path/to/uv",
"args": ["--directory", "/path/to/jadx-mcp-server/", "run", "jadx_mcp_server.py"]
}
}
}zinja-coder maintains it, as part of the Zin MCP Suite, with many community contributors.
Yes, it is Apache-2.0 licensed. JADX itself is a separate open-source decompiler you install yourself.
It works through your MCP client (such as Claude Desktop), so any model costs come from that client's plan or API key.
JADX alone decompiles; this plugin lets an LLM query and reason over the live decompiled context through MCP tools, speeding up review.
Download the jar and server zip from Releases, install the plugin in JADX, set up the server with uv, and add it to your MCP client config.
Exa's MCP server for web search and crawling — give your AI agent a live, high-quality web index.
An MCP server that lets Claude, Cursor and Windsurf build n8n workflows for you
Production-grade MCP server for Godot — let AI build scenes, edit scripts and wire your game live
Native MCP plugin for x64dbg — let an AI drive the debugger: breakpoints, stepping, memory, registers
MCP server that lets AI agents run thousands of Apify Store scrapers and crawlers as tools.
Chrome DevTools for coding agents: an MCP server that lets agents debug and inspect real Chrome
LibreAIカタログに掲載されたAIエージェントプロジェクトの中で最もスターの多いもの — jadx-ai-mcpが合わなかった場合に検討したい、現実的な候補リストです。
A personal AI that really does things — on any OS, with your data staying yours.
An agentic skills framework & software development methodology that works.
Agent harness optimization system with skills, memory and security for Claude Code, Codex and more
The agent that grows with you — a Python framework for adaptive AI agents.
DeepSeek's plugin-based agent harness where everything is a plugin.
カタログの中でjadx-ai-mcpとスター数が近いAIエージェントプロジェクト — 前後の棚もそのまま見て回れます。
Out-of-the-box open-source AI agent for desktop and mobile. Just works.
Production-grade MCP server for Godot — let AI build scenes, edit scripts and wire your game live
Chinese-language course on AI basics: GPU architecture, CUDA, LLMs, and agents.
Real-time voice runtime that keeps AI agents talking, working, and present.
Secure persistent personal agent server in one Rust binary — sandboxed, multi-provider, on your hardware
Open-source observability for AI agents: trace, evaluate, guard, and improve.