Shipping Stargazer before splashdown

Stargazer exists because of a second monitor.

When Artemis II launched, the livestream went up on the spare display and basically never came down. Four people were flying around the Moon and we were in Cincinnati designing interfaces, and every few minutes one of us would drift over to check the stream like it was a fish tank. It was the first time in our lives that deep space felt like something happening now, to people with names, rather than something that happened in documentaries.

Somewhere in that week the question showed up, and it was almost embarrassingly simple: where do you actually look this stuff up? Not the broadcast. The data. Where is the spacecraft right now, what else is up there, what is the Moon doing, what does the neighborhood look like today?

The data was all there, which made it worse

The answer turned out to be that the data is everywhere and the experience is nowhere. NASA publishes constantly. JPL runs an ephemeris service that will tell you where nearly anything in the solar system is at any moment, past or future. ESA, JAXA, and SpaceX all publish their own pieces. It is a staggering amount of public information, and it is scattered across APIs, portals, PDFs, and tools built for people with aerospace degrees.

The existing explorers that do exist are mostly heavy: desktop installs, GPU-hungry simulators, or products priced for institutions. Nothing felt like what we wanted, which was closer to a weather app for space. Open a tab, click a planet, see what is true today, close the tab.

So that became the product: Stargazer, a browser-first space data aggregator with a calm surface. Click a planet, a moon, a spacecraft. Move through the scene. See what it is and where it is right now. Keep the depth available for people who want it, but let the first thirty seconds feel like wonder instead of homework.

A deadline we did not control

Here is the part that made it fun. The mission gave us a clock we could not negotiate with. The crew was coming home in ten days, and we decided the only acceptable version of this project was one that shipped while they were still up there. Building a tribute to a mission after the mission is a memorial. Building it during the mission is participation.

We are a two-person studio. The roadmap meeting was a text message. Scope discipline stopped being a virtue and became a survival requirement, and honestly, the deadline made every decision easier. Feature debates that would normally take a week resolved in seconds, because the answer to almost everything was: does the crew land before we ship this? Then no.

The part where the solar system is too big

The hard problem was never the data plumbing. It was scale. Space is offensively large, and a scene that renders both a spacecraft and the planet it orbits will humble you immediately. Render the solar system honestly and everything interesting is an invisible speck. Compress it enough to see, and you are lying about the very thing that makes space awe-inspiring. Floating point precision starts misbehaving at solar system distances too, which is the kind of bug that does not announce itself so much as make everything gently wrong.

Every space visualization ever made has fought this fight, and like everyone before us we landed on a negotiated settlement: honest positions, merciful presentation, and a camera that cheats on your behalf. I have never respected planetarium software more.

Splashdown

We shipped with the crew still in the capsule. When splashdown finally came, we watched it on the same second monitor the project was born on, with Stargazer open on the first one. I do not have a grander way to describe the feeling than this: the thing on our screen and the thing in the ocean briefly belonged to the same story, and we built one of them.

Stargazer is small and it will grow. But the lesson we keep from it has nothing to do with space. Borrowed deadlines are the best deadlines. Find something real happening in the world, tie your work to it, and let reality do the project management.