The Free Global Intelligence Dashboard Analysts Are Quietly Switching To

If you're an analyst, ops lead, or newsroom researcher who wants a serious situational-awareness dashboard without paying Palantir prices, worldmonitor is likely the first thing you should look at.

It's the largest open-source project in the non-coder-agent ecosystem AgentThread tracks — big enough that it ships as a hosted web app, a desktop app for Mac/Windows/Linux, and as a plug-in your AI agent can query on your behalf. The one thing to know before your organization commits: the license (AGPL v3) is fine for using the tool, but matters if you plan to build on top of it.

Snapshot

What it actually is

worldmonitor is a real-time global intelligence dashboard. Not a news reader. A serious situational-awareness surface with 500+ curated news feeds across 15 categories, a proprietary Country Instability Index (stress-scoring across 31 major countries), an interactive 3D globe plus a flat world map with 56 different data layers, and a finance radar covering 29 stock exchanges plus commodities and crypto.

It ships 6 versions of the dashboard from one codebase — the base world dashboard, plus dedicated tech, finance, commodity, happy (a positive-news filter), and energy versions. Each runs the same engine tuned for a specific reader.

The whole thing can run locally with a free AI model on your own machine (via Ollama), no API keys required. That's important for the analyst whose employer won't let them send data to a third-party AI, and it's rare enough at this scale that it's worth calling out separately.

What ships in the repo

• News + intelligence layer: 500+ curated feeds, AI-summarized into briefs, with cross-references between military, economic, disaster, and escalation stories
• Country Instability Index: stress scoring for 31 major countries with historical trend lines
• Interactive maps: a 3D globe and a flat world map with 56 different data layers
• Finance radar: 29 stock exchanges plus commodities, crypto, and a market-composite signal
• 6 dashboard versions: world, tech, finance, commodity, happy (positive-news), and energy — same engine, different lens
• Native desktop app for macOS (Apple Silicon + Intel), Windows, Linux
• 25 languages including right-to-left scripts
• Agent plug-in so any AI agent (Cowork, Claude Code, ChatGPT Work, Codex, Hermes) can query the dashboard on your behalf via the Smithery MCP server
• Full docs covering architecture, concepts, and self-hosting

For the exhaustive feature list and algorithm documentation, the worldmonitor.app docs are the canonical source.

Why this matters for the non-coder analyst

Three things distinguish worldmonitor from other "AI news dashboard" projects:

First, you can use it however you want. A non-coder can open the hosted web app with zero setup. A more hands-on analyst can install the desktop app. A team can connect their AI agent to it — so "summarize today's escalation signals in Southeast Asia and cross-reference to commodity prices" becomes a plain-English request the agent answers. Very few open-source projects at this scale ship every install path.

Second, the scoring layer is the real product. A generic news aggregator is trivial to build. What worldmonitor actually offers is comparable, scored signals — the Country Instability Index and the cross-referencing between story categories turn raw feeds into numbers you can put in a report. This is the kind of layer a corporate risk team or a newsroom researcher would otherwise pay six figures a year for.

Third, the local-AI option changes who can use it. Legal, defense-adjacent, and healthcare analysts often can't send data to a third-party AI. worldmonitor running with a local AI model on your own machine means the whole workflow — including the AI summarization — stays on your infrastructure. That's the difference between "interesting toy" and "installable in a regulated environment."

Where it fits

Named alternatives in the same broad space:

• Palantir Foundry — commercial, enterprise, six-figure. worldmonitor is not a replacement, but it covers ~60% of the situational-awareness needs of an analyst at 0% of the cost.
• Recorded Future — commercial threat-intelligence platform. Deeper on cybersecurity, thinner on general geopolitical. worldmonitor is broader and cheaper; Recorded Future is deeper on its specific beat.
• Superset — open-source data-dashboard builder. Different job — Superset is a blank canvas for any dashboard; worldmonitor is a ready-built intelligence dashboard with the data already wired in.
• Open WebUI — general chat interface for AI models. If you want "a chat that can browse news," you can wire that up with Open WebUI plus a news plug-in. worldmonitor ships the assembled version.
• claude-obsidian (9,415 stars) — different job (personal note vault) but a comparable "open-source knowledge surface your AI agent can query" for the individual analyst.

Verdict

Install — with the AGPL caveat.

For the individual analyst / researcher: open the hosted web app or install the desktop app. Zero commitment cost, immediate value. If your workflow already runs in Cowork or Claude Code, connect the agent plug-in so your agent can query worldmonitor directly. This is the biggest, most-active project in its category — no reason to look elsewhere first.

For the organization considering embedding this in an internal product: read the AGPL v3 license carefully before you touch it. It's a copyleft license — meaning if you modify the code and expose it over a network (even internally, per some interpretations), you may be required to release your modifications. This is fine for pure use, dangerous for productization without legal review. Consult counsel if in doubt.

Hands-on install-review of the agent plug-in specifically — testing what queries the agent can actually answer and how the results feel in a session — is queued as a follow-up post.

For your agent to install this

You can skip this section — it's the install details your AI agent needs. If you're using Cowork, Claude Code, ChatGPT Work, Codex, or Hermes, say "install worldmonitor for me" and the agent will read the tables below.

System / OS

• Hosted web app (no install): worldmonitor.app — nothing to configure
• Desktop app (Tauri 2): macOS Apple Silicon, macOS Intel, Windows, Linux AppImage
• CLI: npx worldmonitor — requires Node.js 18+
• SDKs: TypeScript (npm i worldmonitor), Python (pip install worldmonitor-sdk), Ruby gem, Go module
• MCP server (agent-facing): worldmonitor/wm-mcp on Smithery — install via any MCP-compatible agent
• Local AI (optional): Ollama installed on the same machine, no additional configuration

API keys — required for any use

None. The hosted app and desktop binary run with zero keys. If you self-host and want AI synthesis, install Ollama locally and no key is needed.

API keys — optional (per specific use case)

Only sign up if the user prefers cloud AI inference over the local Ollama default.

No paid feeds or paid data sources are required — the 500+ news feeds and finance data ship built in.