🌟 NEXUS — Autonomous Multi-Agent AI System (Gemini 4 Argon)

Autonomous multi-agent intelligence system powered by Google Gemini 4 Argon via reverse-engineered cookie authentication, circular multi-account quota rotation, and Model Context Protocol (MCP) support.

✨ Features

• 🧠 Google Gemini 4 Argon Brain: Harnesses the highest-tier reasoning model (gemini-argon) with deep extended thinking, structured planning, and multi-turn conversational memory.

• 🔄 Strict Argon Enforcement & Account Rotation: Automatically detects accounts with Argon capability. If an account lacks Argon permissions or cookies expire, it rotates or reports an explicit error without silently downgrading to Flash.

• 🌐 Real-Time Quota & Account Dashboard: Sleek modern Web GUI (http://127.0.0.1:8000) built with FastAPI, Tailwind CSS, and WebSocket live updates to monitor remaining credits, reset timers, and abuse status.

• 👥 5 Specialist Sub-Agents:

  • CodeAgent: File system operations, script generation, and terminal command execution.

  • BrowserAgent: Headless Chromium automation via Playwright with screenshots and DOM extraction.

  • ResearchAgent: Deep web research via Google Search and DuckDuckGo.

  • MediaAgent: Multimodal asset synthesis, TTS narration (Edge-TTS / ElevenLabs), and poster rendering.

  • YouTubeAgent: Video transcript extraction, comment analysis, and search.

• 🔌 Model Context Protocol (MCP) Server: Full native MCP server (nexus_mcp.py) allowing Antigravity, Claude Desktop, Cursor, or Windsurf to call NEXUS agents and tools seamlessly.

• 🔐 Privacy & Open Source Security: Zero hardcoded secrets. Cookies and credentials stay exclusively in local, git-ignored .env files.

Related MCP server: dat-ai-mcp

🚀 Quick Start

1. Clone the Repository

git clone https://github.com/GanZnosuy/nexus-argon-agent.git
cd nexus-argon-agent

2. Set Up Virtual Environment

python -m venv .venv

# Windows:
.venv\Scripts\activate

# Linux / macOS:
source .venv/bin/activate

3. Install Dependencies

pip install -r requirements.txt
playwright install chromium

4. Configure Environment

Copy .env.example to .env:

copy .env.example .env   # Windows
# or: cp .env.example .env # Linux / Mac

Extract your Gemini Web session cookies (__Secure-1PSID and __Secure-1PSIDTS) from your browser while logged into gemini.google.com, and paste them into .env:

ACTIVE_ACCOUNT=1
ACCOUNT1_SECURE_1PSID=your___Secure-1PSID_here
ACCOUNT1_SECURE_1PSIDTS=your___Secure-1PSIDTS_here

(Optional: Use python quick_cookie.py or run DANG_NHAP_GOOGLE_AI_PRO_NEXUS.bat on Windows for automated cookie capture.)

🖥️ Running the Application

1. Launch the Quota & Account Web Dashboard

python dashboard.py

Open http://127.0.0.1:8000 in your browser to inspect live quotas, switch active accounts, and test Argon reasoning in real time.

2. Interactive CLI Mode

python nexus.py

3. Autonomous Goal Execution

python nexus.py --goal "Research the latest AI breakthroughs this week and compile an executive summary"

4. Setup Model Context Protocol (MCP)

To register NEXUS tools with an MCP host (Antigravity, Claude Desktop, Cursor):

python setup_mcp.py

🛡️ Security & Privacy Notice

This repository is Open Source (MIT).

• Never commit .env or cookies.txt. The .gitignore is pre-configured to strictly exclude all credentials, tokens, session profiles (auth_profiles/), temporary files, and conversation logs.

• All Google authentication tokens remain strictly on your local machine.

📜 License

Distributed under the MIT License. See LICENSE for full details.