Skip to content
ARRANTO
Work
Enterprise case study
Active development
Source on GitHub

Carfax for Boats

Carfax for Boats is still in active development. It is a studio-built project, not a client engagement, and there is no release date yet.

SYSTEM ARCHITECTURE // STUDIO BUILD
STATUS: ACTIVE DEVELOPMENT
PRODUCT ENGINE SPECIFICATION

Carfax for Boats

TECH STACK:Next.jsPWAClaude Vision API

Project overview

Introduction

Carfax for Boats is an Arranto-built project, not a client engagement. It asks a simple question: when a car changes hands, a Carfax report can show its history — what happens when a boat changes hands? The project is under active development, with a public repository showing ongoing commits.

The Problem

A boat's maintenance history typically lives in scattered paper receipts, a seller's verbal account, or nowhere at all. None of it reliably follows the boat when it's sold, which leaves buyers guessing about engine hours, service intervals, and past repairs, and leaves sellers with no easy way to document the upkeep they actually did.

The Technology

The project is built as a Next.js Progressive Web App, so it can be used from a phone at the dock without an app-store install. It uses Claude's vision API to read photos of physical source material — an engine plate, a receipt, an hour meter — rather than requiring manual data entry.

Key Features

  • Photo-Based Input: Capture an engine plate, a receipt, or an hour meter reading directly from the PWA's camera.
  • AI-Assisted Extraction: Claude's vision API reads the photographed details instead of a person typing them in by hand.
  • Structured Records: Extracted information is organized into a maintenance record tied to the boat.
  • Designed to Transfer: Records are built to move with the boat to a new owner rather than staying locked to one account.

Target Use Cases

The project targets boat owners who want a running record of their own maintenance, sellers who want to document a boat's condition before listing it, and buyers who would rather see a structured history than take a seller's word for it.

Challenges

Reading real-world engine plates, faded receipts, and mechanical hour meters is harder than reading a clean digital document — stampings wear down, paper fades, lighting on a boat varies, and handwriting isn't consistent. Vision-based extraction in a marine environment has to work with source material that's often worn, glare-affected, or handwritten rather than printed.

Technology stack
Next.jsPWAClaude Vision API
Repository
Last updated
Aug 7, 2026
Primary language
TypeScript
Source on GitHub

Ready to engineer your custom system?

Let's build resilient, mission-critical infrastructure for your operations.