God's Eye View: The Open-Source Spy-Satellite Globe App

Four days after a solo developer open-sourced a side project, it's sitting at #1 on GitHub's daily trending leaderboard — ahead of every funded AI startup and dev-tool launch trending today. It picked up nearly 2,000 stars today alone, per that same leaderboard. The project looks, at a glance, like a live feed from a spy satellite. It isn't one. It's a browser tab.

That's God's Eye View, released as open source on August 24, 2026 by developer Bilawal Sidhu — a single, 433-file, roughly 220,000-line commit titled "Release God's Eye View as open source". Four days later it had passed 8,000 stars and 1,800 forks. Nothing in the repo, README, or commit history points to a coordinated launch — just a GitHub repo that got shared enough to top the trending page on its own.

What God's Eye View actually is

God's Eye View is a free, MIT-licensed browser app that renders a photorealistic 3D globe — built on CesiumJS and Google's Photorealistic 3D Tiles — and layers live public data on top of it. Per the project's README, the tagline is blunt about the effect it's going for: "A spy-satellite simulator in your browser — then you realize the sources are public and the data is real."

The feature list reads like a mission-control panel: a cockpit view with terrain tracking, a "250 km contact roster," click-to-track on any object with metadata and trails, a voice whiteboard for annotating the map by talking to it, a "scene director" for scripting cinematic camera tours, and a set of sensor filters — CRT, night vision, FLIR/thermal, noir, snow — that reskin the whole globe. It's not a data dashboard; it's built to feel like the UI from a thriller movie, except every layer traces back to a named, public source.

The data behind the effect

Nothing here is exclusive or classified. According to the README, the layers are:

Layer	Source
Live flights (incl. military)	OpenSky + adsb.lol
Live vessels	AISStream
Satellites (~840 objects)	CelesTrak
Earthquakes (last 24h)	USGS
Traffic	TomTom + OpenStreetMap
Public camera feeds (~800)	Cities including Austin, California, London
Active fires (24h)	NASA FIRMS
Radio stations (up to 750)	Radio Browser
Space launches (30-day window)	Launch Library 2
Static infrastructure	4,351 datacenters, 704 dams, 712 submarine cables (bundled)

Every one of those is a public API or open dataset a developer could wire up independently. The product here isn't the data — it's the single UI that makes browsing all of it at once feel like operating a surveillance console.

What it costs to actually run

This isn't a hosted SaaS you sign up for — it's a repo you run yourself (npm install, then npm run dev, on Node.js 24.14.x or 26.x per the README). Most of it is free. The README's own cost breakdown:

• Google Maps (required, for the photorealistic 3D tiles): metered — "the first 1,000 per month are free, then about $6 per 1,000"
• OpenAI (optional, for voice control and the AI HUD): metered, with a built-in "$5 in-app session cap"
• AISStream (live ship data): free, described in the README as "a two-minute signup"
• Flights, satellites, earthquakes, and most other layers: no key required at all

That's a meaningfully different cost model than the commercial flight- and ship-tracking sites this overlaps with in spirit — most of God's Eye View's layers really are $0, and the two paid dependencies are both usage-capped rather than subscriptions.

The feature it explicitly refuses to build

The most interesting design decision in the project isn't a UI feature — it's a line the maintainer drew in the README and is enforcing through the PR process. Quoted directly:

"This project models events, assets, infrastructure, and systems — aircraft, vessels, satellites, fires, cameras, cities. It does not build features for named-person search, face recognition, or tracking individuals, and pull requests that cross that line won't be merged. People are not a query type here."

For a tool that aggregates live flight paths, ship positions, and hundreds of public camera feeds into one browser tab, that's not a throwaway disclaimer — it's the one constraint standing between "open-source OSINT toy" and "open-source stalking tool." Because the project is MIT-licensed, that boundary only holds for the upstream repo; nothing stops a fork from removing it. The README's own safety disclaimer is similarly direct: "Do not use it for flight or maritime navigation, emergency response, medical or health decisions, investment decisions, or other safety-critical or operational purposes. Verify important information with authoritative sources."

Where it's still rough

Four days after a public launch, the seams show. The repo currently lists 53 open issues and 27 open pull requests — a real backlog for a project whose visible maintainer is one person. The README itself is upfront that this is "an evolving open-source client for exploration and learning — a fast, hackable foundation, not a hardened production service." And several of the more interesting layers depend on third-party free tiers (AISStream, NASA FIRMS, TomTom) that a maintainer with no company behind them doesn't control the pricing or uptime of.

The origin story also matters for expectations: per the README, the project comes "from the project behind the viral God's Eye View series (formerly WorldView) — 5M+ on YouTube" — a self-reported view count from a video series, not an audited platform figure. That backstory explains why this shipped polished enough to hit #1 on GitHub within days: it's the open-sourced version of a tool the creator had already been building and using on camera, not a weekend hack.

The takeaway

God's Eye View is a good test case for a pattern worth watching: real public datasets, wrapped in genuinely good 3D UI, released for free under a permissive license, can now out-compete funded products for attention within days of launch, without a marketing budget — GitHub's own trending page did the distribution work that a Product Hunt campaign or ad spend usually does. That's the same distribution shift behind last month's crop of AI-agent browsers, just running through GitHub stars instead of upvotes.

If you're going to run it yourself, budget for the Google Maps key (it's the one layer you can't skip), and read the ethics section before you fork it — the line the maintainer drew there is doing more real work than any of the visual modes. And if you're evaluating your own open-source launch, the more transferable lesson might be smaller: publish the cost breakdown and the "we won't build this" list in the README itself, the way this project did, and let people decide whether to trust it before they've even cloned it — the same falsifiable-claims habit that made Aider's self-reported "percent of code written by AI" metric credible instead of just another marketing number.