LEEWARD Weather Routing

Leeward runs a live pressure field over your fleet and re-solves every voyage as it moves. The same chart your master reads, resolved to a heading.

Western Mediterranean
Field v3 · 0.25° · isobars 4 hPa · Mistral
0 kt 20 40+ kt

A pressure field, not a forecast PDF

Every system on the chart is a vortex with a radius, a spin and a strength. Advect a ship through the sum of them and you have the route. The chart above is drawing exactly that.

LGulf of Lion
996hPa
−2.1 hPa / 3 h · deepening
Barograph24 h
Fig. 01 · SystemsLive

Read the systems

Lows spin one way and draw air in, highs spin the other and push it out. Radius and strength come off the pressure chart; nothing is hand-tuned per voyage.

Under waySOG
  1. Kerkyra Spirit13.4
  2. Bora Trader12.8
  3. Saffron Bay14.1
Fleet fuel, 24 h−3.8%
Fig. 02 · FleetLive

Advect the fleet

Each ship is a particle with a fuel curve and a laycan. We step it through the field, the way the chart steps its 14,000 particles, and keep the tracks that arrive on time.

Next solve0:06
Last diff ETA −0.6 h Course 082° → 076°
Fig. 03 · Re-solveDemo cadence 6 s

Re-solve as it moves

Six-minute cadence. When a system deepens, the route changes with it, and the bridge gets a heading instead of a 40-page routing advice.

One page. Every waypoint.

Each solve prints to a single advisory sheet sized for a 40 kB attachment over Inmarsat. Positions, courses and speeds the officer of the watch can steer, with the wind and pressure the solver saw at each one.

  • PositionsDegrees and minutes, ready to enter into ECDIS. No conversion on the bridge.
  • ConditionsForecast wind and sea-level pressure at every waypoint, so the master can see why the track bends.
  • Re-solveA trigger at the bottom of the sheet says when the next one comes and what would bring it forward.
LEEWARD ROUTING ADVISORY No. 1284-0417 · Kerkyra Spirit · ESALG → GRPIR
Position
36°05′N 005°20′W
Course · speed
076° · 13.6 kt
Wind · MSLP
W 12 kt · 1016
ETA
04 Mar · 06:00Z
Track −7.2 h vs great circle · −21 mt fuel · 1,442 nm Next solve 06:06Z, sooner if any WP moves > 2 hPa

Voyages under routing right now

Colour is wind on the leg, on the same scale as the legend under the chart. Delta is against the great-circle plan the charterer signed.

routing-desk / live-board 6 voyages · agent idle 06:00:00Z
Vessel Leg Agent Wind on route, 24 h ETA delta

One call per voyage

Post the plan, get back a track: waypoints, the field each one was solved against, and the fuel curve that picked it. Everything on the bridge sheet is in the response.

Read the docs

# solve a voyage against the live field
POST https://api.leeward.example/v3/solve
{
  "vessel": "IMO9421268",
  "depart":  "ESALG",
  "arrive":  "GRPIR",
  "speed_kt": 13.5,
  "preset":  "mistral"
}

→ 200
{
  "track":    64 waypoints,
  "eta":      "2026-03-08T17:25Z",
  "fuel_mt":  318.4,
  "delta":    "-7.2h vs great circle"
}

Per hull, per month

No routing-advice fees, no per-call billing. A hull is under routing or it isn't.

Single1 hull

$490

Per hull · month

  • Live field and presets
  • 6-minute re-solve
  • Email and bridge sheet
Operator50+ hulls

Annual

Contract · SLA

  • Private field workers
  • Your own preset library
  • On-prem export of every solve

Put one hull on the field

We open a berth a week. Tell us the vessel and we'll route the next leg alongside your current provider, then show you both tracks.

Where does the field come from?

Global pressure and wind at 0.25°, re-gridded onto our own vortex model every six minutes. The chart on this page runs the same model at demo scale, so you can see what the solver sees.

Is this a forecast or a routing advice?

Both, and neither is the deliverable. The deliverable is a track: waypoints and speeds your bridge can steer, with the field that produced them attached so a master can argue with it.

How does it reach the bridge?

API, email, or the one-page advisory sheet over Inmarsat, which fits in a 40 kB attachment. The board on this page is the same data, rendered for the office.

What happens when the field is wrong?

Every solve is stored with the field it used. When a system deepens faster than modelled, the re-solve arrives with a diff instead of a new PDF that quietly replaces the old one.