Is There a Free AIS API for Vessel Positions? The Honest Answer, and What Actually Moves Ships

Two live records cover the movement question, and they divide the work cleanly.

AISHub Open AIS Data Sharing Network owns the cooperative terrestrial layer: one snapshot per vessel per pull, filterable by bounding box, named vessel list or age of fix, with a separate station-status surface documenting every contributing site. Fixes are live, each stamped with the vessel's last received report, and a maximum-fix-age filter keeps stale positions out of a pull.

VT Explorer AIS Data API is the commercial counterpart, scored 7/10 on the same rubric: 300,000+ vessels tracked daily from a proprietary coastal network extended by satellite reception, positions joined to voyage declarations, ship particulars down to GT, DWT and TEU capacity, expected arrivals, and port-call history queryable over explicit date ranges.

The verified sample tells one whole voyage in three rows - vessel MMSI 304491000 departing Yilport (TRYPO) on 6 August 2017, fixed mid-Mediterranean under way at 7.4 knots with "GIOIA TAURO" declared as destination, then arriving Gioia Tauro (ITGIT) just before 03:00 UTC on the 12th:

# Row 1 — port-call event: DEPARTURE
MMSI         304491000          # vessel identifier, stable across rows
IMO          9175717
EVENT        DEPARTURE
TIMESTAMP    2017-08-06 15:37:22 UTC
PORT         Yilport            # TRYPO — Turkey

# Row 2 — position fix, mid-voyage
NAME         MARENO             # TYPE 70 = cargo
LAT / LON    38.44518 / 15.88136
COURSE       96.7 deg           # SPEED 7.4 kn, HEADING 91
NAVSTAT      0                  # under way using engine
DRAUGHT      8.8 m              # DESTINATION "GIOIA TAURO"
SRC          TER                # TER terrestrial antenna | SAT satellite

# Row 3 — port-call event: ARRIVAL, six days later
MMSI         304491000
EVENT        ARRIVAL
TIMESTAMP    2017-08-12 02:57:05 UTC
PORT         Gioia Tauro        # ITGIT — Italy

Read together they are the analytical move position-only feeds make you assemble by hand: departure geometry, transit kinematics and berthing event already sharing one MMSI key. Weigh the two records field-by-field in the AISHub versus VT Explorer comparison.

Reception physics decides what any AIS record can promise, so name the physics before naming a vendor.

Terrestrial networks - AISHub among them - hear hulls within roughly VHF line-of-sight of an antenna. Coverage hugs coastlines: straits, port approaches and anchorages light up where volunteers cluster, which is precisely where port and supply-chain work happens, while mid-ocean tracks drop out between shores. That makes the network strong exactly at the last fifty miles and quiet everywhere past the horizon.

Satellite reception fills those gaps at coarser refresh, which is why VT Explorer tags each fix by receiving source - TER for a coastal antenna, SAT for an orbital pass - and reports offshore segments the shore layer never sees.

One structural fact outranks all of them: the aggregate moves constantly. The interesting analyses - anchorage density, port-approach queues, fleet behaviour - come from persistence rather than a single pull, and neither record publishes a historical archive of past positions. Your trend lines start accruing from your first delivery, so the earlier the feed lands in your warehouse, the longer the series you own.

Positions alone are dots on water. Three moves convert them into operations-grade records, and each one leans on a named dataset from the same slice.

Buffer a port geography, then test inbound fixes against it. BTS Principal Ports - National Transportation Atlas Database (quality 10) draws 150 principal U.S. port polygons with codes, Coastal/Internal/Great Lakes typing and calendar-year 2023 tonnage; Natural Earth - World Ports (quality 8) adds 1,081 world-port points for global work. Point-in-polygon against those geometries once per polling interval turns a stream of MMSIs into arrivals and departures - unless your feed already returns the event as a row, which is the structural advantage of VT Explorer's port-call records.

Join detected calls to cargo context. USACE Waterborne Commerce Statistics Center (quality 9) publishes annual tonnage, commodity mix and vessel-trip reports for 1,381 locations (819 ports, 562 waterways) across data years 2002-2024, sliced five ways at once: location, commodity, direction, vessel class and draft band. Seasonal rhythm comes from this ledger, not from the live feed.

Mind the clocks before joining anything. AIS moves in minutes, USACE publishes annually, the static layers never change - bridge time grains explicitly rather than joining tables directly, and keep the present-tense feed and the historical ledger as separate layers that share keys instead of pretending one table covers both.

Hulls are half the berth; water is the other half, and one record owns it. NOAA CO-OPS API - Tides, Currents & Water Levels, scored 9/10, observes the physical ocean at 43 PORTS installations: 6-minute water levels with spread and quality flags across 301 active stations, high/low predictions from 3,499 harmonic reference stations, per-depth-bin currents at 86 current stations, plus winds, salinity and air gap - with monthly-mean archives reaching back as far as two centuries at long-running stations.

The sample shows why analysts trust it. Beside every height reading sits a standard deviation folded from the underlying one-second samples; a row that drifts outside the three-sigma band carries the flag in the open instead of silently smoothing itself away:

# observed water level - 6-minute spacing, station 9414290 San Francisco
t : 2026-08-20 00:00    v : 1.825    s : 0.037    f : 0,0,0,0    q : p
t : 2026-08-20 00:06    v : 1.827    s : 0.037    f : 1,0,0,0    q : p   # QC flag tripped

# tide prediction - high/low extreme, station 8518750 The Battery, New York
t : 2026-08-21 01:46    v : 0.417    type : L              # low water

Berth-scheduling logic therefore needs no vendor at all for the water column - the only thing NOAA cannot tell you is which vessel is alongside, which is exactly the question the AIS records answer. Resolve customs handling separately against the CBP Ports of Entry directory (quality 8), which lists all 328 U.S. ports of entry with codes, addresses and office hours, and size national markets against the World Bank panels: container port traffic for 207 economies, 2005-2024, and the Liner Shipping Connectivity Index for 169 economies, 2006-2021.

Six kinds of teams drive most vessel-position searches, and each starts from a different altitude.

Port operations and berth planning watch anchorage occupancy and approach queues build in near real time, sizing waiting time from position history rather than anecdote. Their build patterns live on the developers & builders page for marine ports & services.

Trade-flow and commodity analysts proxy export and import momentum by counting hull movements through chokepoints, weeks before customs statistics publish.

Fleet monitoring and chartering desks track named vessels by MMSI or IMO number with speed, course, draught and destination attached to every fix.

Machine-learning teams train trajectory prediction, port-call estimation and anomaly models on streams that carry navigational status and rate of turn - labelled kinematics most position dumps strip out. Their workflows sit on the data scientists use cases page.

Journalists and academics study shipping behaviour, sanctions-evasion patterns and dark-fleet gaps with a citable source of record.

Supply-chain visibility teams fuse broadcast ETAs into inland logistics planning instead of waiting on carrier notifications.

The pattern underneath all six: the live layer answers where is it now, the registers answer what is this place, and the ledgers answer how much usually moves here. Products get interesting when all three land in one schema keyed on MMSI and port code.

Every figure on this page comes straight from the cataloged source records, verified during the August 2026 research pass - and every claim survives contact with a sample before it survives contact with production.

Request a marine ports & services sample and Datadory ships real rows first: position fixes cut to your bounding box, fleet list and freshness window; port-call events trimmed to your date ranges and UN/LOCODEs; station-status records for the coastlines you monitor; or water-level extracts for the berths you schedule. Name the cut in the request and the sample comes back shaped like your problem - field dictionary and coverage statement attached, yours to keep either way.

Keep going from here:

1. Marine ports & services data hub — all eleven cataloged records with field dictionaries, sample rows and coverage detail.
2. AISHub versus VT Explorer comparison — the cooperative and commercial movement records weighed side by side.
3. How developers & builders use marine ports & services data — turning USACE, BTS and NOAA records into one operational backend.