๐ŸŒ Long View ยท Crucix

AI-powered OSINT intelligence aggregator โ€” 27 data sources, 3D globe, multi-LLM briefings, and real-time delta tracking

Node.js Express Multi-LLM WebGL 3D Globe 27 OSINT Sources Telegram / Discord

๐Ÿ“‹ Project Overview & Problem Statement

Challenge: Staying informed across geopolitics, markets, climate, health, and technology requires monitoring dozens of sources simultaneously. By the time a human analyst reads and synthesizes all relevant signals, the situation has already changed. Traditional news aggregators show headlines but don't surface the signal โ€” they lack synthesis, delta tracking, and intelligent alerting.

Solution: Long View (codenamed Crucix) is a local intelligence aggregator that pulls data from 27 OSINT sources every 15 minutes, detects meaningful changes using a delta engine, generates AI briefings using whichever LLM the operator specifies (Claude, OpenAI, Gemini, or Codex), renders the data on a 3D WebGL globe, and pushes alerts to Telegram and Discord when significant shifts occur.

Key Capabilities

๐Ÿ–ฅ๏ธ Architecture & Components

๐Ÿ“ก Source Aggregator

27 standalone source modules in apis/sources/. Each exports a briefing() function that returns structured data. Sources degrade gracefully when API keys are absent โ€” the full sweep continues even if individual sources fail.

๐Ÿ”„ Delta Engine

Compares each new sweep against the previous state. Uses configurable numeric thresholds per source (e.g., VIX must move ยฑ3 to trigger alert). Memory stored in lib/delta/memory.mjs โ€” persists across restarts.

๐Ÿค– Multi-LLM Briefing

Provider-agnostic LLM layer in lib/llm/provider.mjs. Set LLM_PROVIDER to anthropic, openai, gemini, or codex in the environment. Generates synthesized narrative briefings from aggregated source data.

๐ŸŒ 3D Globe Dashboard

WebGL globe rendered via Three.js showing geographic signal distribution. The Jarvis-style HUD displays current intelligence indicators. Designed for ambient monitoring โ€” always-on display gives a quick situational overview.

๐Ÿ“ฒ Alert Bots

Telegram bot (polling-based) and Discord bot (webhook or full bot with slash commands) push delta alerts when the engine flags a significant change. Configurable per source โ€” customize which signals trigger alerts.

๐Ÿณ Docker Support

Includes docker-compose.yml for containerized deployment. Run the entire stack โ€” aggregator, dashboard, and bots โ€” with a single docker compose up command.

๐Ÿ“Š Key Metrics

27
OSINT Sources
4
LLM Providers Supported
15 min
Default Refresh Cycle
2
Alert Channels (TG + Discord)

๐Ÿ“ก Intelligence Source Categories

Financial Markets Geopolitical Risk Climate & Weather Health Surveillance Social Sentiment Satellite Imaging Conflict Monitoring Supply Chain Energy Markets Commodity Prices Cyber Threat Intel Economic Indicators

๐Ÿ› ๏ธ Tech Stack & Architecture

Node.js 22+ Express Claude API OpenAI API Google Gemini API WebGL / Three.js Telegram Bot API Discord Bot API Docker Pure ESM

Quick Start

# 1. Copy .env.example to .env # Add your LLM API key and bot tokens LLM_PROVIDER=anthropic LLM_API_KEY=sk-ant-your-key-here TELEGRAM_BOT_TOKEN=your-telegram-bot-token DISCORD_WEBHOOK_URL=your-discord-webhook # 2. Start with Docker (recommended) docker compose up # OR run directly node server.mjs # Dashboard at http://localhost:3117