Data, not a promise
Observations and forecasts help compare spots but cannot guarantee a session. The 0–96 index is a transparent comparison heuristic, not a probability or a command to launch.
AN NPO NT AI PROJECT FOR RIDERS
About Kite + AI
An AI-assisted planner built for Saint Petersburg riders: forecasts, sensors, cameras and spot parameters brought together for today, tomorrow and the next seven days.
The shore matters
Every spot has a working wind sector. Strong offshore wind receives a lower result regardless of calculated kite size.
A person decides
The service cannot see every obstacle, local squall or equipment issue. Riders and instructors assess the real conditions at the shore.
AN INITIATIVE OF ARTEM POLTORATSKIY
Want to ride? Ride. Do not waste time comparing sensors.
LLC “SPU Scientific Technologies” (NPO NT) develops AI projects for biotechnology and medicine. Kite + AI is an open applied experiment in using the same methods to collect heterogeneous information, check its consistency and turn it into a concise rider summary.
The portal is designed for people who work, ride in their free time and should not need professional-level weather analysis. It reduces repetitive information requests to instructors and station operators; instructors remain essential for teaching, shore safety and local assistance.
The north-western loop currently includes 10 physical stations and 11 observation channels that could be connected reliably from public sources. We do not call this every sensor in existence. The transparent model combines wind, gusts, a shore-relative working sector, agreement between three forecast models, precipitation, observation freshness, forecast horizon, rider weight and level, and the selected kite.
The next technical stage is an archive of forecast-to-observation pairs. Once a forecast horizon closes, the system will be able to compare an earlier prediction with observed values and estimate systematic error by model, station and area. The public version rechecks data availability and freshness every minute, but does not present that process as automatic neural-network retraining.
Public streams may contain stale readings, persistent zero values, duplicate-channel conflicts or stations that poorly represent beach wind. Without a technical inspection we cannot claim a particular installation violates a standard. AI can flag anomalies and reduce the weight of a doubtful signal, but it cannot reconstruct true wind from a broken measurement.
Use the portal in real conditions and tell us how well the AI handles the task — practical feedback is the point of this experiment.
AN OPEN PORTAL
Instructors and kite stations.
We are planning an open directory where instructors, schools and stations can describe their work and region. The first step is to understand whether riders and professionals need it — without hidden promotion or an implied qualification check by the portal.
To join the future directory or suggest its format, contact us on Telegram. Nothing will be published before a separate agreement.
Message @mikahelp →RIDER FEEDBACK
Help the AI become more useful.
Tell us where you ride and what is clear or needs improvement. No name, phone number, email address or account is required.
A review appears immediately after automated technical checks. Do not include personal data, contacts, links, advertising or information about other people. To request removal, message @mikahelp with the review text and date.
What riders say
0 reviewsLoading reviews…
DATA SOURCES