OSINT dashboard

Israel Monitor is a public OSINT dashboard for monitoring the Israel-Iran conflict — aggregating live rocket sirens, AI-generated briefings, social media feeds, flight tracking, and prediction markets into a single military-grade command interface. As the sole designer and developer, I owned everything: product vision, UI/UX, data architecture, and infrastructure — using it as a deliberate experiment in vibe-coding and self-teaching cloud services, live data fetching, and backend automation.

2026

VISIT

18
Live data sources

Infrastructure

70+
Monitored accounts

Signal curation

5-min
Alert refresh rate

Response time

730+
AI briefings delivered

Intelligence

Israel Monitor Real-Time OSINT dashboard-2.png

Context

No single place to understand what was happening, right now

When the Israel-Iran conflict escalated, existing sources couldn't keep up. News sites lagged hours behind. Twitter was noise. Telegram channels broadcast Hebrew sirens to a mostly Israeli audience. There was no English-language, real-time tool that combined threat level, context, and signal in one place — so I built one.

Israel Monitor was built to fill that gap: a public, free, real-time intelligence dashboard that aggregates sirens, news, social media, flight data, prediction markets, and AI-generated briefings into one focused command-center interface.

What We Built

A command center for a conflict zone

Israel Monitor is a single-page React application with a dark, military-grade design aesthetic, built around three core pillars:

Real-time situational awareness — A live Red Alert system scrapes Israel's official Home Front Command (Pikud HaOref) every 5 minutes, maps incoming rocket and drone alerts to an interactive Leaflet map, groups them by geographic region (translated from Hebrew to English), and triggers audio alerts in the browser. A radar SVG animation pulses when sirens are active, goes still when it's quiet.

Intelligence & context — Twice daily, an AI model synthesizes recent headlines, timeline events, and social media posts into a structured Morning and Evening Briefing: a strategic recap plus five "Things to Watch" items with likelihood and impact ratings — automatically pushed to the @israelmonitor Telegram channel at 07:00 and 20:00 IST, and exportable as Instagram Story images directly from the dashboard.

Signal over noise — A curated feed of 70+ monitored accounts (IDF, CENTCOM, OSINT analysts, journalists, think tanks) surfaces the posts that matter. A breaking ticker auto-scrolls critical headlines weighted by source priority. Three live charts track news volume, siren activity, and internet connectivity for both Israel and Iran via Cloudflare Radar — useful for detecting infrastructure disruptions before they hit the news cycle.

The full stack: React 18 + TypeScript + Vite, Supabase (PostgreSQL + Deno edge functions), 18 serverless functions, pg_cron scheduling, Recharts, and Leaflet.js.

Early vibed design result

Fully redesigned UI

This project was also a deliberate exercise in vibe-coding and self-teaching. I used Lovable as the primary build environment, but pushed well beyond prompting — writing a Node.js relay server from scratch to bypass CORS on the Pikud HaOref API, learning how HLS video streams are structured to embed live TV, configuring cron jobs on Supabase, and parsing and deduplicating RSS feeds at scale. Every feature was a new layer of understanding, built under real-world pressure, with real data, during a real conflict.

OUTCOME

Shipped and used during an active conflict

Israel Monitor went live at israelmonitor.org and became a reference tool during escalation periods. The Telegram channel grew organically as a reliable source for English-language structured briefings. The project proved that a solo designer-developer — armed with modern tooling (Lovable, Supabase, AI models) and a willingness to learn in public — can ship production-grade, data-dense infrastructure fast.

The real takeaway wasn't the product. It was the compounding confidence of building something that works under real-world pressure, with real data, during real events.