TL;DR: We built something nobody has. A live 3D globe showing exactly where your code executes on Cloudflare's edge network. 2.7KB of Zig WASM. 17 cities. 4 continents. Real data. Watch it live β
The question nobody answers
There are tools that tell you how many times your code ran. Dashboards that show error rates, latency percentiles, and CPU time. Datadog charges $50,000 a year for them.
But none of them answer the simplest question: where does my code actually run?
Not which AWS region. Not which availability zone. Where β on a spinning 3D globe, with real dots, in real time.
We checked. Not Cloudflare. Not Datadog. Not Vercel. Not AWS. Nobody.
So we built it.
The numbers
Less than 24 hours since launch:
Atlanta. Paris. Amsterdam. London. New Delhi. Singapore. Hong Kong. Stockholm. Warsaw. And more lighting up every hour.
How it works
Every request to mcpaas.live runs a 2.7KB Zig WASM binary on the nearest Cloudflare edge. That's the FAF scoring kernel β compiled from Zig, targeting WASM MVP, no WASI, no bulk-memory. It scores AI context in sub-millisecond time.
When a request hits, one thing happens in the background:
const colo = request.cf?.colo; // 'ATL', 'CDG', 'LHR'ctx.waitUntil(Β Β env.KV.get(`globe:${colo}`).then(v =>Β Β Β Β env.KV.put(`globe:${colo}`, String(Number(v || 0) + 1))Β Β ));One KV read, one KV write, non-blocking. The request isn't slowed. The colo code (IATA airport code β LAX, LHR, NRT) tells us exactly which Cloudflare data center processed the request.
Two layers, one story
The globe shows two types of dots:
- Cyan dots β every Cloudflare edge location. 175 mapped. The network. The potential. "We can run here."
- Orange dots β locations with real executions. Sized by count. Growing in real time. "We are running here."
Auto-rotates. Responds to mouse and touch. Keyboard controls (β β Space). Auto-refreshes every 60 seconds. On load, randomly picks a cinematic entry β zooming out from the busiest dot, or swooping in from deep space.
Why Zig?
The WASM binary is 2,702 bytes. For comparison:
Zig gets this small because of comptime evaluation β most of the work happens
at compile time. The runtime binary is just the hot path. No allocator in the scoring loop,
no WASI overhead, WASM MVP target only.
Same spec, three implementations, three orders of magnitude difference.
The stack (costs $0/month)
- Compute: Cloudflare Workers (free tier)
- Storage: Cloudflare KV (free tier)
- Binary: 2.7KB Zig WASM
- Visualization: globe.gl from CDN
- Coordinates: 175 IATA locations in a TypeScript file
- Third-party services: None
No database. No Redis. No analytics vendor. The entire infrastructure costs nothing.
The product question
The observability market is $3.35 billion and growing at 15% annually. Datadog makes $2.45 billion a year. Grafana crossed $400 million ARR. Teams spend $50,000+ per month on monitoring β and their monitoring costs now exceed their compute costs.
But none of them show you a globe.
We're asking ourselves: is "see where your code runs" a feature or a product? Right now it's a feature of MCPaaS. But the reaction in the first 24 hours suggests it might be more.
If you run code on Cloudflare Workers, Vercel Edge Functions, or any edge runtime β wouldn't you want to see this for your own code?
The leaderboard
C'mon Rio. Sydney. Tokyo. Got no devs?
Put your city on the map
The globe is live. The data is collecting. Every MCP connection, every namepoint claim, every soul page visit β it all shows up as a dot.
Visit mcpaas.live/globe and watch your city light up.