MCP-1st-Birthday/fleetmind-in-action

<div align="center">

🚛 FleetMind AI Agent

Autonomous Fleet Dispatcher Powered by Gemini 2.0 Flash

The Problem: Managing delivery dispatch requires constant decision-making—who to assign, which route to take, how to handle urgent orders.

The Solution: FleetMind AI Agent autonomously handles complex dispatch workflows through natural conversation, powered by Gemini 2.0 Flash and 29 MCP tools.

The Impact: Say "Create urgent delivery for Sarah, find the best driver considering weather" and watch AI plan, reason, and execute the entire workflow.

     

</div>

📺 See It In Action

<div align="center"> <a href="https://www.youtube.com/watch?v=chIWtFA20dU"> <img src="https://img.youtube.com/vi/chIWtFA20dU/maxresdefault.jpg" alt="FleetMind AI Agent Demo" width="560"> </a> <p><b>▶️ Click to watch the demo video</b></p> </div>

Watch the autonomous AI agent plan, reason, and execute multi-step delivery dispatch workflows using Gemini 2.0 Flash and 29 MCP tools.

⚡ Quick Start

Step 1: Open the App

👉 [Launch FleetMind AI Agent](https://huggingface.co/spaces/MCP-1st-Birthday/fleetmind-in-action)

Step 2: Connect to MCP Server

• —Enter MCP Server URL (pre-filled)
• —Enter your API Key (Get one here)
• —Click Connect

Step 3: Start Dispatching!

You: "Create an urgent delivery for Dr. Emily Watson at UCSF Medical Center,
     fragile medical supplies, assign the best driver"

Agent:
  🔧 Step 1: Geocoding address... ✓ (37.7631, -122.4586)
  🔧 Step 2: Creating order... ✓ ORD-abc123
  🔧 Step 3: AI assignment... ✓ Driver Mike selected (92% confidence)

  ✅ Created urgent order for Dr. Watson and assigned to Mike Chen.
     ETA: 25 minutes. AI selected Mike because: closest driver with
     refrigerated van, experienced with medical supplies.

🏆 What Makes This Special?

🛠️ The 29 Tools

<details> <summary><b>📍 Geocoding & Routing (3)</b></summary>

• —geocode_address - Address → GPS coordinates
• —calculate_route - Traffic-aware routing
• —calculate_intelligent_route - Weather + traffic AI routing

</details>

<details> <summary><b>📦 Orders (8)</b></summary>

• —create_order, fetch_orders, get_order_details
• —search_orders, get_incomplete_orders, count_orders
• —update_order, delete_order

</details>

<details> <summary><b>👥 Drivers (8)</b></summary>

• —create_driver, fetch_drivers, get_driver_details
• —search_drivers, get_available_drivers, count_drivers
• —update_driver, delete_driver

</details>

<details> <summary><b>🔗 Assignments (8)</b></summary>

• —create_assignment - Manual assignment
• —`auto_assign_order` - Nearest driver
• —`intelligent_assign_order` - 🤖 Gemini AI assignment
• —get_assignment_details, update_assignment
• —unassign_order, complete_delivery, fail_delivery

</details>

<details> <summary><b>🗑️ Bulk (2)</b></summary>

• —delete_all_orders, delete_all_drivers

</details>

🏗️ Architecture

┌─────────────────────────────────────────────────────────┐
│                   Gradio UI (app.py)                    │
│      Dark theme • Reasoning visualization • Chat        │
└────────────────────────┬────────────────────────────────┘
                         │
                         ▼
┌─────────────────────────────────────────────────────────┐
│              FleetMind Agent (agent.py)                 │
│  ┌───────────────┐  ┌───────────────┐  ┌─────────────┐ │
│  │    Context    │  │ Gemini 2.0    │  │  Iterative  │ │
│  │  Engineering  │  │    Flash      │  │ Agentic Loop│ │
│  └───────────────┘  └───────────────┘  └─────────────┘ │
└────────────────────────┬────────────────────────────────┘
                         │ MCP SDK (SSE)
                         ▼
┌─────────────────────────────────────────────────────────┐
│           FleetMind MCP Server (29 tools)               │
│   PostgreSQL • Google Maps • Gemini AI • OpenWeather    │
└─────────────────────────────────────────────────────────┘

🚀 Local Development

# Clone & install
git clone https://github.com/mashrur-rahman-fahim/fleetmind-agent-track2.git
cd fleetmind-agent-track2
pip install -r requirements.txt

# Configure .env
MCP_SERVER_URL=https://mcp-1st-birthday-fleetmind-dispatch-ai.hf.space
MCP_API_KEY=fm_your_key_here
GEMINI_API_KEY=your_gemini_key_here

# Run
python app.py

📊 Technical Specs

🔗 Links

👥 Team

FleetMind Development Team

MCP 1st Birthday Hackathon - Track 2: MCP in Action (Enterprise Category)

<div align="center">

Built with ❤️ using Gemini 2.0 Flash + MCP for the MCP 1st Birthday Hackathon

  

</div>