There is no shortage of free wind charts and GRIB data. However, this still leaves the key questions for trip planning unanswered: Will the planned route work under sail? Will we need to tack? How long will the crossing realistically take? And is it worth setting sail a few hours later?
This is exactly where OhMyWind comes in. The French open-source project combines weather and sea conditions with a simplified speed calculation based on boat polars. The route can be created and calculated directly in the web interface. There, you can also set a time window to compare several departure or arrival times. Unfortunately, the web interface is only available in French, but it can be used via the translation function built into many web browsers. Alternatively, the software can be linked to an MCP-compatible AI client. This translates a query formulated in everyday language into structured calculation requests for OhMyWind and interprets and explains the results.
OhMyWind therefore fulfils a different role to the comprehensive weather routing applications that YACHT covers in its overview of Weather and Routing Apps 2026 Note: This is a freely accessible route planner, not a complete on-board navigation system.
Weather models by region and forecast period
OhMyWind does not produce its own forecasts. According to the Project methodology on GitHub the application retrieves wind data via the forecast interface from Open-Meteo It varies and switches between several models depending on the area and time frame.
For the French Atlantic coast and the Mediterranean, Météo-France’s AROME model is intended to have a grid size of 1.3 kilometres. The high-resolution model covers approximately 48 hours and is designed to better capture local thermal effects as well as the Mistral and Tramontane. For Europe, the German Weather Service’s ICON-EU model is used, with a grid size of seven kilometres and a forecast period of around five days. For the medium and longer term, OhMyWind uses the ECMWF IFS and, as a fallback, the GFS.
The model cascade is a useful approach. No single model can provide high resolution, a long lead time and global coverage all at once. However, it does not replace the skipper’s own assessment of the weather conditions. Particularly in unstable conditions, a calculated route should not be taken as a precise forecast for every bay or headland.
Route, polar curves and a realistically reduced pace
For a particular leg of a voyage, OhMyWind divides the route into sections and calculates the wind angle, expected boat speed, estimated time and warnings for each section. At the current stage of the project, seven polar archetypes are available. These range from a 20-foot cruising boat to a 50-foot blue-water yacht, as well as racer-cruisers and a 40-foot catamaran. A specific boat type is not automatically derived from a shipyard database. However, the AI client can take over this task and select the best polar curve based on length and description.
The underlying polars are deliberately not intended to calculate best-case times. By default, OhMyWind applies an efficiency factor of 0.75. This takes into account reductions caused by trim, comfort requirements, the helmsman, load and other factors. If the destination is to windward, the software takes into account tacking via the VMG. The course shown on the chart remains a straight line between the start and the destination. OhMyWind therefore does not display actual tacking legs. The additional time required is based on a mathematically reduced speed in the direction of the destination. This serves as a useful initial guide when deciding whether a route is even feasible in headwinds. However, it is not sufficient for actual route planning.
This distinction is important. Even paid apps differ considerably in the level of detail with which they calculate and display routes. The Practical test of Navimetrix shows just how far professional systems have now come in terms of GRIB analysis, routing and navigation.
Sea state and currents: More accurate in France than elsewhere
In addition to wind data, OhMyWind utilises data on sea state, surface currents and water levels from Open-Meteo’s marine interface. This global data has a grid size of eight kilometres. It is suitable for rough route planning on the open sea and in the Mediterranean, but does not show narrow fairways or local current edges in the level of detail found on a local chart.
On the French Atlantic coast, the project also draws on MARC Atlas data from Previmer, Ifremer and SHOM. High-resolution data is available there for tidal passages, for example in the Finistère region, the English Channel, southern Brittany and Aquitaine. However, the methodology itself also highlights a limitation: the data captures the tidal component, but not short-term weather-induced changes in currents.
According to the developers, tides and currents in the Mediterranean are deliberately not taken into account in the current version. For a crossing from Marseille to Porquerolles, OhMyWind can therefore assess the wind and sea state, but cannot provide a local current forecast. In any case, when sailing in tidal waters, passage planning is no substitute for a current atlas or local knowledge of the area.
Compare departure times rather than just working out the next departure
The project’s most powerful computational feature is the ability to simulate multiple departure times. The ‘Compare Windows’ tab is available in the web interface for this purpose. The user sets the start and end of the period, selects the interval (every hour, every three hours or every six hours) and specifies the boat type. OhMyWind then calculates the individual routes. Hourly departure windows can be set up to 14 days in advance, giving a maximum of 336 variations. For each of these, the service provides times for each section of the route, wind angle, polar speed, warnings and a comfort rating from one to five.
The comparison can also be initiated via a connected MCP-compatible AI client. There, the user specifies the route, boat and time slot in the chat. The client passes the data on to OhMyWind and can then summarise the returned options verbally. Without an AI client, OhMyWind does not determine which time slot is best, nor does it make any recommendations. This interpretative task falls to the AI client or to the user via the web interface.
Also Windy’s Route Planner uses multiple comparisons. However, YACHT has shown that a well-presented sail analysis is of little help if the calculated route itself is implausible. OhMyWind solves this problem differently: it does not automatically optimise the route at all. The sailor or the AI client sets the waypoints. This ensures that responsibility for ensuring the route makes sense clearly remains with the user.
AI Dialogue: Useful, but not a green light to let it run its course
OhMyWind can be connected to AI clients such as Claude, ChatGPT or other supported applications via the Model Context Protocol. A query might be: “I’d like to sail from Gelting to Lyö on Friday in my X-3.” The AI client assigns the boat in question to one of the seven Polar archetypes and passes the route, time and boat type to OhMyWind.
OhMyWind then calculates the route and returns structured data: times for each section of the route, wind angle, polar speed, warnings, complexity score and a link to the map view. If the client supports the MCP Apps specification, it can display this interactive OhMyWind view directly within the chat. Otherwise, it summarises the values provided as text and links to the web view.
The assessment is only carried out within the AI client. It interprets the data, compares the calculated options where there are multiple start times, and is able to articulate the pros and cons of a time slot in natural language.
No nautical chart, no automatic alternative route
The biggest practical limitation lies with the map. OhMyWind is a planning tool, not an electronic nautical chart. Depths, fairways, restricted areas, traffic separation schemes and local hazards must be checked against up-to-date, official chart data before routing. The application is therefore unable to determine whether a direct route between two waypoints is navigable.
Similarly, there is currently no iterative high-precision calculation or automatic route optimisation. The time calculation is carried out as a single run. If the calculated arrival time changes, the weather conditions along the route are not repeatedly updated until the time and forecast have converged. Coastal effects at headlands, jet effects and detours to avoid unfavourable wind fields must be modelled using custom waypoints.
That’s no small matter, but a genuine limitation. The comparison with PredictWind and its routing functions shows just how much extra work and costly infrastructure can be involved in more comprehensive solutions. OhMyWind does not wish to replicate this range of functions at present.
Book recommendations
Sailing in tidal waters explains how tides and tidal currents can be taken into account when planning specific sailing trips. The focus is on sources of information, technical tools, methods for determining tides and typical problems in coastal areas. This complements the functionality of OhMyWind, particularly where the app utilises high-resolution data for French tidal waters, though it does not replace a traditional tidal atlas.
Weather at sea provides a more comprehensive overview of weather observation and forecasting. Among other things, the book covers local and regional weather events, sea states, ocean currents, meteorological trip planning and weather-dependent sailing. For readers who wish not merely to accept a weather routing result but to verify it using their own assessment of the weather situation and local conditions, it provides the necessary fundamentals.
In your view, what tasks should a free weather routing service fulfil, and where do you think the responsibilities begin that no software can take on?
Hauke Schmidt
Test & Technology editor