Skip to main content
Engineering · August 6, 2026 · 8 min read

Five products, one maritime spine

Binnacle started as crew credential tracking and turned into five products — fleet compliance, a pilot association dispatch board, harbor management, USCG exam prep, and an ARPA radar simulator. Why they share a data model, and why the regulated middle of an industry is a better place to build than the consumer edge.

Container ship alongside gantry cranes at a commercial port

A binnacle is the stand that holds a ship’s compass. It is bolted to the deck in front of the helm, at eye level, and its whole job is to keep the one instrument you steer by lit, level, and where you can see it. It is not glamorous hardware. Nobody photographs the binnacle.

That is roughly the product thesis.

Binnacle began as one thing — crew credential tracking for commercial fleet operators — and is now five products sharing a spine. This is a note on how that happened, what actually got built, and why the boring regulated middle of an industry turns out to be a better place to build software than the consumer edge everyone competes over.

The problem is that the paperwork never stops

Commercial maritime is a business where an expired document stops a vessel.

A licensed mariner carries a Merchant Mariner Credential with endorsements, each with its own expiry. Layered on top: a TWIC card, a medical certificate, drug-testing enrollment under 49 CFR Part 40, STCW basic training with five-year refreshers, firefighting, first aid, radar observer. A fleet operator running twenty vessels is tracking hundreds of expiry dates belonging to people who are, by the nature of the work, frequently not in the office.

Then the regulatory surface on the vessel side: Subchapter M inspection and work/rest hour limits for towing vessels. STCW 77/10 rest hours for the international trade. Drill and training records that a Coast Guard inspector will ask to see. CG-2692 marine casualty reports, due within five days of an incident and generally written by someone who has just had a very bad day. Drydock and survey schedules. Insurance certificates. ISM safety-management document control. ISPS and MARSEC visitor logs. OFAC screening on counterparties. Since 2025, cyber compliance under 46 CFR 101 Subpart F, which surprised a lot of operators who did not previously consider themselves to be in the information-security business.

Almost none of this is tracked in software today. It is tracked in a spreadsheet maintained by one person who has been there fifteen years, and a wall calendar, and that person’s memory. It works until they retire, or until an inspector arrives on the wrong Tuesday.

That is the wedge. Not “AI for maritime.” A list of expiry dates that nobody wants to keep by hand, attached to consequences that are immediate and financial.

Where the second, third, fourth, and fifth products came from

None of these were roadmap items. Every one came from noticing that the same spine — vessels, people, credentials, time windows, and an auditable record of who knew what when — already fit an adjacent problem that a different set of people were solving with a whiteboard.

Binnacle Passage is a live dispatch board for US marine pilot associations. Pilots board inbound ships and con them into port; the association coordinating them needs to know which vessels are inbound, their ETA to the boarding ground, who is next in the rotation, and whether that pilot has legal rest hours available. It wants live AIS coverage, departure and launch alerts, digital order intake, under-keel clearance and tidal-window planning, tariff invoicing, and a sealed pilotage record afterward.

Read that list against the fleet product. Vessels with positions and ETAs. People with credentials and rest-hour limits. Time windows with hard constraints. Sealed records for later dispute. It is the same object graph pointed at a different desk. The genuinely novel part was turn-equity — a rotation is not just a queue, because a pilot who drew three short jobs is owed a long one, and the association’s internal politics live in that math.

Binnacle Harbor came from the other end of the size scale. A small-boat harbor runs on a paper logbook: slip assignments by hand, billing calculated at checkout with a calculator, problems discovered during a walk-through rather than when they happen, and storm prep communicated by phone tree. Harbor puts the whole thing on one screen — a live slip map rendering real-time occupancy, vessel check-in in about thirty seconds, nightly rate times nights stayed with balances that surface on their own, an issue log with severity alerts, and one button that sends a storm-precaution checklist to every slip holder at once. There is a live demo running on Kona Small Boat Harbor.

The pricing follows the customer rather than the software: it is free to run your harbor, and the harbormaster keeps 100% of every payment, with boaters covering a small card fee. Small harbors do not have a software budget line. They have a card reader and a filing cabinet.

Binnacle School exists because the compliance product kept surfacing the same upstream problem: people need the credential in the first place. It is USCG deck exam prep — 2,418 questions across twelve categories, every single one with a written explanation, scaling by credential from 1,027 for the OUPV/Six-Pack up through the 100-Ton and 200-Ton Master pools. Three subjects are free with explanations plus one full timed exam, because a study app that will not let you study before paying is just a paywall wearing a hat. It syncs between web and iOS, and the AI tutor is grounded in exam material rather than answering from open-ended guesswork — with the standing caveat, stated on the page, that it is a study aid and anything safety-critical gets confirmed against the source.

It is also, explicitly and in writing, not affiliated with or endorsed by the Coast Guard. In a regulated space that disclaimer is not legal boilerplate to be minimized. It is a trust signal, and burying it would cost more than it saves.

Binnacle Radar is the strangest of the five and the most fun to build: an ARPA bridge radar simulator running COLREGS scenarios. Range rings, gain and sea and rain clutter controls, EBLs and VRMs, target acquisition with CPA and TCPA limits, relative and true motion, head-up and north-up, trial maneuvers, and a helm with a telegraph. Radar observer endorsements require real time in front of a real display, and the equipment costs what shipboard equipment costs. A faithful simulator is not a replacement for that hour. It is the twenty hours before it, which currently do not exist for most candidates.

Why the regulated middle

The consumer edge of maritime — boat rentals, marina finders, fishing apps — is crowded, low-priced, and seasonal. The regulated middle has properties that are unusual and, from a builder’s perspective, extremely attractive:

The deadline is not negotiable. An expired credential does not degrade the user experience. It stops the vessel. Software that prevents that has an obvious price relative to a day of lost revenue.

The rules are public and stable. 46 CFR, 49 CFR Part 40, Subchapter M, STCW — these are published, citable, and change on a legislative timescale rather than a competitor’s release schedule. You can encode them. They will still be true next quarter.

The incumbent is a spreadsheet. Not a well-funded competitor with a sales team. A spreadsheet, and the person maintaining it usually knows it is a liability.

Nobody wants to build here. It requires reading a lot of federal register text and understanding why a towing vessel’s rest hours differ from a deep-sea watchstander’s. That is a real moat, and it is made entirely of tedium.

The tradeoff is honest and worth stating: this is not a market you win with a good landing page. Buyers are conservative, sales cycles run long, and a demo that gets a regulation subtly wrong ends the conversation on the spot. You have to actually know the domain. Building it from Honolulu — a working harbor town where the customers are down the road and half of them can tell you which pier they are tied up at — helps considerably.

The spine

The thing holding five products together is not a shared UI kit. It is a single model of vessels, people, credentials, time windows, and an auditable record of state changes — who knew what, and when.

That last part is the load-bearing one. In every one of these products, the valuable artifact is not the current state. It is the defensible record of the state at a moment in the past, produced when an inspector, an insurer, an attorney, or a class society asks. A fleet operator proving a drill happened. A pilot association proving rest hours were legal at the time of the transit. A harbormaster proving a storm warning went out before the weather did. A mariner proving a credential was current on the day of the voyage.

Build that once, carefully, with immutability where it matters, and the products on top of it are mostly a question of which desk you are pointing it at.

The compass has to stay lit. That is the whole job.

Tagged
maritime compliance USCG Hawaii software product strategy
Begin

Building something of consequence?

Begin a design engagement →