NEW LOCATIONS ADDED 11/07/2005 - TIDE AND CURRENT DATA NOT FUNCTIONAL ( Version 1.2.1 - alpha)
The Waterbike.ai Origin Story: Safe by Design
How a would-be waterbiker took 8 years of experiences and trained an AI to read the Bay so riders can explore with confidence, have fun, and return safely.
Scene: A Bay that does not suffer guesswork
San Francisco Bay is gorgeous and moody. It rewards preparation and humbles shortcuts. We respect the inherent risk. In the past, “careful” meant too many tabs, too much guessing, and not enough clarity. New riders asked, “Is today good?” The honest answer was, “It depends, give me twenty minutes.”
The decision: Build the guide we always wanted
We started Waterbike.ai with one rule: augment human judgment, do not replace it. Guides make the final call. The system gathers the right facts, explains why they matter, and keeps us safe by default.
We coded the model around three truths:
The Bay is a system. Wind, tide, current, daylight, topography, and traffic interact.
Routes are not monolithic. Each leg has a different personality and exposure.
Safety is personal. A strong intermediate can enjoy conditions that a new rider should defer.
The build: curiosity meets code
We gave the model senses with live marine data, a brain with routing and scoring logic, and a conscience with clear rules that choose the safer path when uncertain. There were lessons. We fixed a time drift that confused tide queries. We learned the API’s hilo setting to get true high and low events. We split the rulebook from the reasoning so guides can tune thresholds without shipping code. A daily job now wakes up, computes, publishes a small plan file, and goes back to sleep. No drama.
We also show our work. Every green or yellow badge comes with human-readable reasons.
What makes this different
• Leg-aware currents. We project current as a vector on each route leg to estimate real effort and exposure.
• Skill-aware recommendations. Choose Casual, Intermediate, or Expert. Thresholds adjust to match ability and fitness.
• Windows, not guesses. We compute when to go, not just whether to go, aligned to daylight and tide behavior.
• Explainable by design. Stations, thresholds, and fallbacks are listed so a human can validate the call.
• Fast and dependable. A static plan file powers the site. No spinner. No mystery.
The payoff: confidence and time back
For instructors, this replaces scattered research with one defensible brief, saving hundreds of hours each season. For members and guests, it delivers calm confidence. Open the page and see today through the next 4 to 5 days, the likely wind, the current mood, safe windows, and an honest effort score that matches skill and fitness. That helps gear choices, launch timing, and group safety so everyone can return safely.
A day in the life
Pick Pier 40 to Tiburon with a short layover. The system proposes a start that uses a friendly ebb out of South Beach, a manageable flood near Raccoon Strait, and a daylight buffer on return. It flags a ferry window to avoid and suggests sheltered bailouts. The summary reads: Effort 6 out of 10, Intermediate, wind 9 to 14 mph with gusts to 18, current up to 1.2 knots with you outbound. A short note explains why this window keeps the group safer.
Culture: safe by design
We do not sell bravado. We provide clarity. The Bay will always carry inherent risk. That is part of its beauty. With Waterbike.ai, guides move faster from data to decision, riders feel capable instead of lucky, and shared adventures become repeatable and teachable so groups return safely.
What is next
Effort heatmaps per leg. Shadow corridors that reveal wind funnels and lee zones. Smart layovers that turn tide flips into a rest break. Gentle alerts when a new green window appears. Same philosophy. More useful ways to say: go now, go prepared, and return safely.
This is just the beginning, actually Day 3!
This began as one rider’s curiosity and became a daily promise to the community. We will do the hard reading so you can do the fun riding. The Bay is still complex and alive. With Waterbike.ai, it is more understandable, and that understanding helps keep us safe.
Waterbike.ai : Timeline & Sprint Calendar
(Nov 2025–Jan 2026)
Cadence: 2‑week sprints (with a 4‑day
Roles: Damien (Product/Founder), Echo (Front End Eng), GIA (QA/Automation/Backend), Guides (Operator Feedback)
Backend Milestone Map
M0 : Hardening & Setup (Nov 6–9, 2025): Repos, environment, time‑zone fix, NOAA interval=hilo, baseline leg maps.
M1 : Core Daily Plan v1 (Nov 10–23): Deterministic daily plan.json for the 4 routes, automated publish, website embed (Today).
M2 : Explainability & Skill Profiles (Nov 24–Dec 7): Green/Yellow/Red with reasons; Casual/Intermediate/Expert thresholds; leg‑aware current projections; docs.
M3 : 5‑Day Horizon & Instructor Brief (Dec 8–21): Rolling 5‑day plan, caching, instructor‑ready summary, simple alerts.
M4 : Launch & Talk Prep (Dec 22–Jan 4, 2026): Slide deck, demo scenes, rehearsal clips; soft launch + partner feedback.
Features
Core Backend Engine (The "Nautilus" Planner)
Automated Data Ingestion: Automatically fetches live marine advisories, wind data, and tide/current predictions from multiple NOAA sources.
Centralized Rulebook: All safety thresholds for wind, gusts, current, and daylight are managed in a simple, external configuration file (thresholds.yaml).
Multi-Level Safety Checks:
Applies absolute "No-Go" rules (like Small Craft Advisories) that halt all planning.
Enforces hard, pod-specific limits based on rider skill levels (e.g., "Beginner Pod" vs. "Pro Pod").
Intelligent Tide Window Solver: Analyzes high and low tide events to identify optimal windows of time that minimize adverse currents.
Smart Route Generation: Uses a graph-based algorithm (like A*) to dynamically find the best and safest routes between waypoints, preferring safe crossings and efficient paths.
Operational Logic: Automatically calculates staffing requirements (escort, sweep) and maximum group size for each recommended trip.
Comprehensive Safety Score: Generates a single, easy-to-understand safety score (0-100) and a color-coded badge (Green, Yellow, Red) for every possible ride.
Explainable AI: Provides simple, human-readable reasons (why) for its recommendations and lists any specific rules that were triggered (blockers).
Automated Publishing: Runs on a daily schedule via GitHub Actions and automatically publishes its findings as simple data files (windows.csv, plan.json) for apps to consume.
Architectural Relationship
This is the key: your new systems are consumers of the core system we are building now.
+------------------+ +----------------------+
| |------->| |
| Neptune App | | |
| (Fitness Coach) | | Nautilus Engine |
| | | (Prediction API) |
+------------------+ | |
^ | |
| +----------------------+
| ^
| |
+------------------+ |
| |----------------------+
| Olympus App |
| (AR Game) |
| |
+------------------+
Wish List
Future User-Facing Apps
Echo: The User-Facing Planner App (Phase 5)
Multi-Platform Support: Designed to work seamlessly on web, native mobile (iOS/Android), and tablets.
Offline Capability: Allows users to view the day's pre-generated plan even without an internet connection.
Personalized Experience: Tailors ETA and effort scores based on an individual rider's past performance.
Live On-Water Assistance:
Provides real-time rerouting suggestions if a rider deviates from the planned course.
Triggers geofenced alerts for dangerous areas like ferry lanes or known chop zones.
Interactive Checklists: Delivers on-water checklists and landmark cues to improve safety and navigation.
Feedback Loop: Gathers post-ride feedback from users ("Was it spicy?") to continuously learn from real-world conditions and improve the AI's future recommendations.
Neptune: The Fitness Coach App (Phase 6 - Future Product)
Advanced User Profiling: Securely stores user biometrics including height, weight, sex, BMR, HRV, VO2max, and resting heart rate.
Wearable Integration: Connects with popular health platforms (Apple HealthKit, Garmin, etc.) to sync real-world workout data.
Predictive Health Impact: Utilizes a proprietary algorithm to forecast the physiological impact of a planned ride, including estimated calorimetry, Metabolic Equivalents (METs), and heart rate zone duration.
Performance Analytics: Tracks user performance over time, correlating it with environmental conditions to provide actionable insights for training.
Multi-Sport Support: Adapts its models and recommendations for various water sports (kayaking, SUP, e-foiling) based on different physiological demands.
Olympus: The AR Watersports Game App (Phase 7 Future Product)
Real-Time Pod Racing: Allows users to join a virtual "pod" and compete against friends in real-time on the water.
Augmented Reality (AR) Courses: Overlays "virtual buoys" and racecourses onto a user's view, creating an immersive gaming experience.
Live Leaderboards: Displays real-time rankings during a race and maintains historical leaderboards for specific routes.
Intelligent Handicapping: Automatically generates handicaps for mixed races based on user skill level, equipment type, and live environmental data from the Nautilus engine.
Safety-First Design: Integrates directly with the Nautilus API to ensure all virtual events and races operate within the established safety parameters for the day's conditions.