World Monitor – AI‑powered global intelligence dashboard

What it is – A TypeScript‑based web‑app (and Tauri desktop app) that aggregates news, geopolitical, financial, energy, and infrastructure data into a single, real‑time situational‑awareness interface. The platform uses local LLM inference (Ollama) to synthesize briefings and exposes its data through a programmatic API that agents can call.

---

Core features (as described in the README)

Feature	What you get
Curated AI‑summarized news feeds	Global and regional news streams turned into short briefs by an on‑device LLM.
Dual map engine	3‑D globe (globe.gl) and flat WebGL map (deck.gl) sharing a common layer catalog.
Specialised panels	Separate UI variants for world, tech, finance, commodity, happy, and energy domains, all built from the same codebase.
Cross‑stream correlation	Signals from military, economic, disaster, and escalation data are correlated to surface emerging risks.
Country Instability Index (CII)	Live v8 scores for 31 Tier‑1 countries, updated roughly every 24 h.
Finance radar	Real‑time feeds for stock exchanges, commodities, crypto and a market‑wide composite index.
Local AI inference	Runs entirely with Ollama (no external API keys), using Transformers.js in the browser.
Multilingual UI	Native‑language feeds, RTL support, and translations for the interface.
Desktop app	Tauri 2‑based native binaries for Windows, macOS (Intel & Apple Silicon) and Linux (AppImage).
Programmatic access	• MCP streaming server (/mcp) with tools list & call endpoints (auth via header or OAuth). • OpenAPI‑described REST API (api.worldmonitor.app). • Official CLI (worldmonitor / wm). • SDKs for Python, Ruby, Go (zero‑dependency).

---

Architecture snapshot (from the README)

• Frontend – Vanilla TypeScript, Vite, globe.gl + Three.js, deck.gl + MapLibre GL.
• Desktop – Tauri 2 (Rust) with a Node.js side‑car.
• AI/ML – Ollama for local LLM inference; Transformers.js for browser‑side model execution.
• API contracts – Protocol Buffers with sebuf HTTP annotations.
• Deployment – Vercel Edge Functions, Railway relay, static PWA, and Tauri binaries.
• Caching – Upstash Redis, three‑tier cache hierarchy, CDN, service‑worker.

---

Getting started (quick‑start)

# clone, install, run dev server
git clone https://github.com/koala7/worldmonitor.git
cd worldmonitor
npm install
npm run dev          # http://localhost:3000

Variant‑specific commands (e.g., npm run dev:finance) spin up the same app with a different domain configuration. A self‑hosting guide covers Vercel, Docker, or static deployment.

---

How to use it programmatically

• MCP server – https://worldmonitor.app/mcp streams tool definitions. Example:

  npx worldmonitor tools            # list tools (no key needed)
  worldmonitor risk IR --api-key wm_…
  

• REST API – OpenAPI spec at https://worldmonitor.app/openapi.yaml.
• SDKs – Install the language‑specific client (pip install worldmonitor-sdk, gem install worldmonitor, go get …/sdk/go).
• API keys – Obtain a key at worldmonitor.app/pro for authenticated calls.

---

License & commercial use

Source code is AGPL‑3.0‑only. Commercial SaaS deployments are allowed if the AGPL obligations are met, or a separate commercial licence can be negotiated.

---

Who’s behind it

Created and maintained by Elie Habib (GitHub @koala73) with community contributors. Security disclosures are handled via a public security policy.

---

Bottom line

World Monitor is a full‑stack, AI‑enhanced dashboard that merges real‑time geopolitical, financial, and infrastructure data into an interactive map‑based UI, while also offering a rich set of APIs and SDKs for agents and scripts. Its emphasis on local LLM inference and open‑source licensing makes it a noteworthy example of an AI‑driven situational‑awareness platform.