FTC Dashboard

A comprehensive club management platform for FIRST Tech Challenge (FTC) teams, featuring real-time attendance tracking, task management, budget oversight, and AI-powered insights. Created by teams #10937 and #30548 Stuttgart High School.

Installation

You can install the FTC Dashboard globally via npm:

npm install -g ftc-dashboard

Or run it directly using npx:

npx ftc-dashboard

Features

• Dashboard: Live club health metrics, attendance trends, and activity summaries
• Teams & Members: Manage team rosters with role-based permissions and scopes
• Attendance: Grid-based attendance tracking with AI-powered analysis
• Tasks: Kanban board for task management with completion analytics
• Budget: Income and expense tracking with financial summaries
• Outreach: Log community service hours and events
• Communications: Track emails and announcements
• Messaging: In-app chat with mention notifications
• AI Scout: Real-time FTC news and updates powered by local LLM
• WebSocket: Live updates across all connected clients

Tech Stack

• Frontend: React + TypeScript, Tailwind CSS, Framer Motion
• Backend: Express.js with SQLite database
• AI: node-llama-cpp with local GGUF models (e.g., Phi-3.5)
• Real-time: WebSocket for live notifications and chat

Prerequisites

• Node.js 18+
• SQLite3 (included with better-sqlite3)
• A GGUF model file (Phi-3.5 recommended, ~3GB)

Quick Start

As a CLI tool

Simply run the dashboard from your terminal:

npx ftc-dashboard

Or, for installation:

npm install ftc-dashboard

This will guide you through the setup on your first run.

Local Development

1. Install dependencies:

   npm install

2. Start the development server:

   npm run dev

   The app will be available at http://localhost:3000

Configuration

Environment Variables

Create a .env file in the project root:

LLAMA_MODEL_PATH=./models/Phi-3.5-mini-instruct-Q5_K_M.gguf
MAX_TOKENS_LIMIT=1024
DISABLE_NEWS=0
OLLAMA_URL=http://localhost:11434/api/generate
OLLAMA_MODEL=phi3.5
EXA_API_KEY=your_exa_api_key_here

• MAX_TOKENS_LIMIT: Maximum tokens for AI responses (default: 1024)
• DISABLE_NEWS: Set to 1 to disable the news endpoint temporarily
• OLLAMA_*: For Ollama backend integration (optional)
• EXA_API_KEY: For web search integration (optional)

API Endpoints

AI Endpoints

• POST /api/ai/fetch-news - Get FTC/robotics news (supports ?stream=true)
• POST /api/ai/attendance - Analyze attendance patterns
• POST /api/ai/check-excuse - Evaluate absence reasons
• POST /api/ai/activity-summary - Generate club activity summaries

Data Endpoints

• GET/POST /api/teams, /api/members, /api/attendance, /api/tasks, /api/budget, /api/outreach, /api/communications
• POST /api/auth/login, /api/auth/setup, /api/auth/reset

Model Performance

The app uses CPU-only inference by default to avoid GPU allocation errors. For optimal performance:

• Use quantized models (Q4_K_M, Q5_K_M) to reduce memory footprint
• Phi-3.5 mini (~3-4GB) provides good balance of speed and quality
• Response generation typically takes 30-60 seconds depending on model size and hardware

Database

SQLite database (nexus.db) stores:

• Teams and members with role-based permissions
• Attendance records with excused/unexcused tracking
• Tasks with status tracking and assignments
• Budget transactions categorized by team
• Outreach events and hours logged
• Messages and notifications with real-time WebSocket sync

Development

• Build for production: npm run build
• Preview production build: npm run preview
• Hot Module Replacement enabled for instant feedback

License

GNU General Public License v3.0