⚡ SCéAL.AI: Momentum Tracker Agent ⚡

SCéAL.AI is an AI-powered analytics engine that monitors project activity across multiple channels — GitHub, Twitter, Onchain, and Community (Discord) — and distills it into real-time momentum scores, anomaly alerts, and AI-generated insights.

Built for founders, investors, analysts, and ecosystem explorers who want to track Web3 project activity and signals in one unified intelligence dashboard.

---

✨ Features

• 🧠 AI Insights from GitHub, Onchain, and Discord data via Groq + LLaMA
• 📊 Momentum Scoring across GitHub, Twitter, Onchain, and Community streams
• 🔎 Anomaly Detection & time-series trend analysis
• 💡 Insights Dashboard with breakdown cards, sparkline visualizations, and highlights
• 🔌 Modular Fetchers with real APIs (GitHub, Etherscan, Uniswap, Discord)
• 🧱 Built with Next.js App Router, Groq SDK, Tailwind CSS, and Recharts

---

🧱 Architecture Overview

.
├── app/
│   ├── github/               → GitHub Analytics page
│   ├── onchain/              → Onchain Token Analytics page
│   ├── discord/              → Discord Activity Analytics
│   └── components/           → Visualizations, charts, cards
│
├── lib/
│   ├── agents/               → Momentum Agent core & time-series logic
│   ├── data/                 → Fetchers: GitHub, Twitter, Onchain, Discord
│   ├── ai/                   → AI Insight generators using Groq + Vercel SDK
│   └── scoring/              → Scoring logic, anomaly detection, trend tracking
│
├── types/                    → Shared types: `MomentumData`, `MomentumScore`, etc.
├── public/                   → Assets and visual elements
└── api/                      → Routes for GitHub, Discord, and Onchain metrics

⚙️ Setup Instructions

1. Clone the repository

git clone https://github.com/YOUR_USERNAME/momentum-tracker.git
cd momentum-tracker

2. Install the dependencies

# Using pnpm
pnpm install

# or npm
npm install

3. Create a .env.local file

GITHUB_TOKEN=github-api-key
TWITTER_BEARER_TOKEN=twitter-dev-api-key
ETHERSCAN_API_KEY=api-key
ALCHEMY_RPC_URL=alchemy-rpc-url
GROQ_API_KEY=your-groq-api-key
DISCORD_TOKEN=your-discord-token

Note: Ensure that the discord bot has enough privileged permissions to read a channel's messages, otherwise it may lead to fault during data fetching.

4. Run the dev server

npm run dev
# or
pnpm dev

Open localhost:3000 to get started

📊 Data Streams & Insights

Each tab uses specialized fetchers + LLaMA models to summarize activity into clear insights.

🟣 GitHub

• Metrics: Stars, Forks, PRs, Issues, Velocity, Contributors
• AI: Health summary + growth suggestions

🟢 Onchain

• Metrics: Volume, Transactions, Holders, DEX Pairs, Liquidity
• AI: Token flow and liquidity risk analysis

🔵 Discord

• Metrics: Messages, Active Users, Message Length, Top Contributors
• AI: Community health + engagement anomalies

🔍 Anomaly & Trend Detection

Momentum Tracker uses:

• 📈 Moving averages for smoothed time-series
• ⚠️ Trend change detection via average deltas
• 🔔 Spike alerting with variance & threshold logic
• ⛓️ Federated breakdowns: Discord + Onchain + GitHub in unified cards

🧪 Modules

• MomentumAgent – Central orchestrator
• TimeSeriesAnalyzer – EMA, trend detection, federated breakdown
• AnomalyDetector – Spike & pattern detection
• AI Insights – Groq-generated text over GitHub, Onchain, Discord data
• GitHubDataFetcher – REST API for velocity, PRs, contributors
• OnchainDataFetcher – Etherscan + DeFiLlama + Dexscreener
• DiscordDataFetcher – Reads server messages with bot token

🧠 AI-Powered Insight Modules

We use Groq's LLaMA-3 to generate structured object insights:

{
  summary: string;
  outlook: "bullish" | "bearish" | "neutral";
  keySignals: string[];
  riskLevel: "low" | "medium" | "high";
  reason: string;
}

Supported generators:

• generateGitHubInsight(metrics)
• generateOnchainInsight(metrics)
• generateDiscordInsight(stats)

📈 Visualizations

• Recharts for AreaCharts (Time Series)
• Framer Motion for animated cards
• Lucide for iconography
• Tailwind CSS for glassmorphism UI
• Next.js App Router for isolated tab routing

📊 Evaluation & Diagrams

Below is an example evaluation breakdown (optional to include as a Jupyter notebook):

Metric Type	Inputs	Aggregation Engine	Output
GitHub	stars, forks, commits, PRs	TimeSeriesAnalyzer	GitHub Momentum Score
Onchain	tx count, holders, liquidity	Weighted Aggregation	Onchain Activity Score
Discord	messages, user counts	AI + Stats	Community Health Score
RAG	All combined data	LLM (Groq)	AI-generated Insights + Alerts

🚀 Future Enhancements

• 🧩 Telegram & Reddit community integration
• 📥 Wallet-based custom momentum dashboards
• 🔔 Real-time notifications on momentum spikes
• 📊 CSV/JSON data export for research
• 🔌 Plugin architecture for data pipelines

🙌 Credits & Acknowledgments

This project would not be possible without the contributions and technology from:

Core Libraries

• Next.js
• TailwindCSS
• Framer Motion
• Lucide Icons
• Recharts
• React ScrollArea
• zod

AI & RAG

• Groq — LLM inference on llama3-8b/70b
• AI SDK — AI integration framework

Data Providers

• GitHub API
• Etherscan API
• DeFiLlama
• Dexscreener
• Discord API
• Nitter / Puppeteer Scrapers for Twitter alternatives

Visualization

• Particles.js
• Gradient Hero Backgrounds

Built with ❤️ for the OnlyFounders AI Hackathon