OceanScrape reconstructs vessel positions from publicly visible map tiles across the world's busiest shipping corridors. The maritime supply chain moves roughly 80 percent of global trade. The signals describing it already exist in the open spectrum. Structured access to them does not.
AIS was specified as an open broadcast. Its modern aggregation is anything but. Insurers, defense analysts, journalists and researchers routinely pay for access to data that was, by design, public. OceanScrape documents what is already visible on free renderers and produces a geometric dataset: counts, vessel types, motion states, attributed by region and time.
OceanScrape is built on a small set of explicit commitments. They constrain what the tool does, who it serves, and how its data leaves the system.
Shipping lanes are infrastructure. The volume and rhythm of traffic through them shapes prices, conflicts and ecology for every coastline. Knowing what moves through the commons should not be a paid subscription.
The dataset is reconstructed from what cameras, satellites and terrestrial receivers already see. No upstream feed is trusted on its word. Every count traces back to a pixel and a timestamp.
The detection model, region geometry, and database schema are visible in the repository. A claim about maritime activity is only as credible as the code that produced it. Black boxes are not evidence.
Output is geometric and aggregate. Counts, types, motion states. The tool does not identify individuals, crews, or beneficial owners, and is not engineered to. Capability is bounded on purpose.
Four stages, each reproducible from the source tree. No proprietary feed, no licensed AIS aggregator, no broker contract.