moonloop

Our thesis

We should conquer the oceans.

Most of the seafloor has never been properly mapped. That gets said a lot. Here’s the part that doesn’t: much of the ocean we have surveyed is just as invisible, because the data is sitting in a proprietary file on a drive in an office that one program can open and one person understands.

The people doing this work are very good at it. They’ll hold a vehicle to centimetres in a current, four hundred metres down. The software around them is the weak part. Every sonar manufacturer writes its own format. Every autonomy stack writes its own log. A dive comes back as five or six files that don’t talk to each other, and the first hour of review goes on lining them up by hand.

So people share screenshots. Not because they want to — because a screenshot is the only thing that reliably opens on someone else’s computer.

Then there’s timing. On land you watch a system while it runs. Underwater, the vehicle goes in, everyone waits, and the review happens after recovery. By then the ship has moved, and the line that needed reflying costs a day instead of ten minutes.

None of this is really anyone’s fault. Software was an accessory to selling sonar. If the hardware is the business, your file format is a feature, not a public good. That held for thirty years and it worked well enough.

We think it’s over. Offshore wind, subsea cable, defence and science are all scaling at once. Autonomous vehicles got cheap enough to put several in the water at a time. And a browser can now render a whole mission without an install. The bottleneck stopped being how much you can record and became how much you can actually look at.

So we’re building three things, in this order.

Drop any underwater data file into a browser and see the dive straight away — sonar, track, telemetry, video, one timeline. No install, no licence, and a link anyone can open without an account. Free, permanently. That isn’t a loss leader, it’s how the thing spreads.

Then the workspace people actually live in: mission libraries, 3D bathymetry, annotations and review threads, comparing this quarter against last, deliverables that go to a client as a session instead of a folder.

And underneath both, a vendor-neutral layer that takes data from anything, keeps it where you need it — cloud, private, or air-gapped — streams telemetry while the vehicle is still down, and lets you search the whole archive instead of one project at a time.

Neutral is the point, not a slogan. A layer that works for hardware makers, autonomy teams, survey contractors and hydrographic offices at once only exists if none of them has to lose for it to work. Open specs, real APIs, and export that actually works — including out of us.

Every push into the sea came down to an instrument. The chart, the chronometer, the echo sounder, satellite positioning. Each one turned guesswork into work, and whoever held the instrument got there first. This time the instrument is software.

The ocean isn’t unknowable. It’s unobserved. That’s a much smaller problem, and it’s a software one.