Flowstate announced the release of a new AI model update known as the Surfing Foundation Model, alongside the launch of the Flowstate MCP (Model Context Protocol) integration platform for premium partners. The update follows the processing of 20 million waves through its platform across five wave-making technologies and includes data collected from surf parks and competitive surfing events.
The new model is designed to understand a complete wave ride from take-off to exit and has been trained using proprietary data from surf park venues in six countries, five wave-making technologies, and competition footage gathered through partnerships with the World Surf League (WSL) and US Board Riders.
“The latest advances in the Flowstate AI model moves way beyond detecting objects and classifying actions, but to comprehend a wave ride as a complete event: from take-off to exit, move by move, in the context of where the surfer is on the wave and how that ride compares to every other ride the model has ever seen,” said Luke Wallace, Founder & CEO of Flowstate.
“That’s what makes everything next possible. Automated scoring, AI coaching, gamification and progression aren’t features we’re bolting on, they’re the natural output of a model that has watched, learned from, and understood 20 million waves. The new era of customer experience in surf parks is starting now.”
The model combines manoeuvre detection, wave and manoeuvre scoring, wave positioning feedback and biomechanical feedback into a single system capable of automatically scoring rides without human input or venue-specific configuration.
Flowstate said training data includes footage collected during the Australian leg of the 2026 WSL Championship Tour, including sessions at the Rip Curl Pro Bells Beach and Gold Coast Pro Snapper Rocks. Additional training data was collected through US Board Riders competitions at locations including Trestles and Steamer Lane.
The company added that four AI-powered products are being developed on top of the model and are scheduled for rollout to venue partners during 2026. These include AI gamification and challenges, automated competition scoring, AI coaching and feedback, and AI progression tracking.
“Getting a model to understand surfing isn’t just about recognising manoeuvres,” said Chris Hausler, CTO of Flowstate. “It has to understand the entire ride in context — how sections connect, where moves happen on the wave, and how performance changes over time. That’s what the Flowstate AI model does. Everything we build next starts from there.”
Flowstate also announced that its platform has processed 20 million waves through its video capture, tracking and delivery pipeline. The company said the dataset contributes to the training of the AI model.
The newly launched Flowstate MCP is available to premium partners and provides a standardized integration layer for connecting third-party applications with Flowstate data, rider profiles, scoring outputs and AI tools.
Surf park venue operators can use the MCP to build custom applications including leaderboards, coaching tools, booking integrations, loyalty programs and operational dashboards. Flowstate also released a surfer-facing MCP integration that allows users to connect their accounts to compatible AI assistants for session analysis, coaching insights, progression tracking and performance statistics.
The latest Flowstate AI model update is available exclusively to Flowstate partners. You can find Flowstate at the following wave pools, Palm Springs Surf Club, Waco Surf, Atlantic Park Surf, Revel Surf, Lost Shore Surf Resort, SURFTOWN Munich, The Wave Bristol, URBNSURF Melbourne, URBNSURF Sydney, and many others.
WavePoolMag has followed the evolution of Flowstate for several years, starting with a technical podcast with programmers Chris and Mike. CEO and founder Luke Wallace has consistently given WavePoolMag detailed updates as the surf park space evolves and Flowstate along with it.
