Spaces:
Running
on
Zero
A newer version of the Gradio SDK is available:
6.1.0
title: NACC - Network Agentic Command Control
emoji: 🚀
colorFrom: blue
colorTo: purple
sdk: gradio
app_file: app.py
pinned: false
license: mit
short_description: AI-Powered Network Orchestration via MCP
tags:
- mcp-in-action-track-enterprise
- mcp-in-action-track-consumer
- mcp-in-action-track-creative
- building-mcp-track-enterprise
- building-mcp-track-consumer
- building-mcp-track-creative
🚀 NACC: Network Agentic Command Control
AI-powered orchestration for distributed systems using Model Context Protocol (MCP)
NACC transforms multiple computers into a unified intelligent network, controllable through natural language. Built from scratch for the HF MCP Birthday Hackathon 2025.
🎥 Demo & Social
📺 Watch Full Demo Video - See NACC controlling Mac and Kali Linux in action!
💼 LinkedIn Post - Read about the journey and challenges
🏆 Hackathon Tracks & Justification
We are submitting NACC for all 3 tracks because it represents a fundamental shift in how MCP can be applied across different domains.
### 🏢 Enterprise Track (mcp-in-action-track-enterprise, building-mcp-track-enterprise)
Why it fits:
- Infrastructure Orchestration: NACC solves a massive enterprise pain point - managing distributed servers and VMs. Instead of manual SSH, DevOps teams can use natural language to orchestrate complex workflows across hundreds of nodes.
- Security First: Built with a custom secure handshake protocol and whitelisted command execution, designed for sensitive corporate networks.
- Scalability: The hub-and-spoke architecture allows unlimited nodes to be added dynamically, perfect for large-scale server farms.
### 🏠 Consumer Track (mcp-in-action-track-consumer, building-mcp-track-consumer)
Why it fits:
- Home Automation for Geeks: Imagine controlling your home media server, Raspberry Pi, and gaming rig from a single chat interface. "Check if my Plex server is running" or "Update my Pi-hole" becomes a simple text message.
- Accessibility: Lowers the barrier to entry for managing home labs. You don't need to know complex Linux commands; you just need to know what you want to do.
### 🎨 Creative Track (mcp-in-action-track-creative, building-mcp-track-creative)
Why it fits:
- Novel Architecture: We pioneered a "Two-Space Architecture" (Gradio + Docker) connected via a custom HTTP bridge to simulate a real network within HuggingFace's constrained environment.
- Agentic Innovation: NACC isn't just a chatbot; it's an autonomous agent that plans, routes, and executes tasks across physical boundaries. It "thinks" in network topology.
- Built from Scratch: We didn't use any pre-existing MCP SDKs. We built the entire MCP protocol implementation (JSON-RPC, tool registry, handshake) from zero to demonstrate deep understanding and creative application of the spec.
🎯 HuggingFace Space Demo
Quick Start
- List nodes:
list nodes - Switch to VM:
switch to vm-node-01 - Explore files:
list files→navigate to /app/src - Try file ops:
navigate to /tmp→create a file named demo.txt with content Hello!
Demo Instructions
📚 Full demo guide available in the UI - Open "Help & Instructions" accordion at the bottom
Key Points:
- VM Space (
vm-node-01) - Remote node simulation - File operations work in
/tmp/directory - Use absolute paths for navigation:
/app/src,/tmp - AI-powered natural language commands
🏆 For Judges
What Makes NACC Special
Built from Zero: Every line of code written from scratch. No pre-made MCPs, no templates - pure original implementation.
Original Vision: NACC was designed for real network environments - controlling multiple PCs, servers, and VMs on the same network with full system access.
Space Adaptation: To showcase on HuggingFace Spaces, I created a unique two-space architecture:
- Main Space: Orchestrator + UI (this space)
- VM Space: Simulated remote node
- HTTP Bridge: Connects both spaces to simulate network communication
Trade-offs: Limited operations to protect running code. Real NACC (see demo videos) has full filesystem access, file sharing, and multi-node orchestration.
Features:
- ✅ Natural language command interpretation
- ✅ AI-powered tool selection and routing
- ✅ Cross-space file operations
- ✅ System monitoring and node management
- ✅ MCP protocol implementation
💬 To The Community
Thank you for checking out NACC! This project represents months of learning and experimentation with AI agents and distributed systems.
About Me: Vasanthadithya - 3rd year B.E. CSE student at CBIT, Hyderabad
Looking For: Internship opportunities, learning experiences, and chances to work on real-world AI/security projects.
If you like this approach to network orchestration, let's connect! I'm passionate about building secure, autonomous systems from the ground up.
📧 Reach out: LinkedIn | GitHub
🛠️ Architecture
Two-Space Setup
┌──────────────────────────┐
│ Main Space (This) │
│ - UI + Orchestrator │
│ - AI Backend (Blaxel) │
│ - hf-space-local node │
└────────┬─────────────────┘
│ HTTP
↓
┌──────────────────────────┐
│ VM Space │
│ - NACC Node Server │
│ - vm-node-01 │
│ - Isolated filesystem │
└──────────────────────────┘
Real NACC (Local Network)
┌────────────┐ ┌────────────┐
│ MacBook │ │ Kali Linux │
│ (Node) │ │ (Node) │
└─────┬──────┘ └──────┬─────┘
│ │
└─────────┬──────────┘
│
┌──────┴──────┐
│ Orchestrator│
│ (AI Brain) │
└─────────────┘
🤖 Tech Stack
- Frontend: Gradio
- Backend: FastAPI
- AI: Blaxel (GPT-4o-mini) + fallback regex
- Protocol: MCP (Model Context Protocol)
- Communication: HTTP/REST
- Language: Python 3.10+
📺 Links
- GitHub: Vasanthadithya-mundrathi/NACC
- VM Space: NACC-VM
- Creator: @Vasanthfeb13
Built with ❤️ for HF MCP Birthday Hackathon 2025 🎂