In the world of tactical cycling, information isn't just power, it's safety. One week into the development of BikeScout, we are moving beyond simple data aggregation. We are building a system that understands the physics of the ground under your tires.
Traditional AI is "frozen" in its training data. It can tell you what a mountain is, but it can't tell you if the trail on that mountain is a swamp today. We chose the Model Context Protocol (MCP) because it acts as a real-time "brain extension."
Deterministic vs. Probabilistic
LLMs are probabilistic, they guess. Physics is deterministic. By using MCP, the BikeScout server handles the hard math locally (slopes, rain accumulation, tire pressure) and provides the AI with a "Ground Truth" briefing. The AI then synthesizes this into human advice.
The Mud Engine: 72H-RAIN Analysis
Our current "Mud Risk" tool is the first step toward true predictive intelligence. It's not just about current rain; it's about cumulative saturation.
mud_index = total_rain_72h * soil_sensitivity
# Clay: 1.2 | Gravel: 0.3 | Asphalt: 0.0
Is this simple? Yes. But it's effective. By cross-referencing rainfall volume with soil permeability (clay vs. sand), our engine predicts the "Mud Risk." We are already working to evolve this into a Geophysical Intelligence model, factoring in Evapotranspiration (how fast wind and sun dry the trail).
Vertical Reality (SRTM-V3)
Generalist apps "smooth" elevation, hiding 20% gradient spikes. BikeScout uses NASA SRTM-V3 30m data. If a trail has a 22% gradient on loose gravel, BikeScout flags it as a "Power Drain" for e-bikes or a "Hike-a-Bike" for gravel grinders. We don't just show a line; we show the effort.
Dynamic Tire Intelligence
The most recent push to our repository addresses the most critical rider decision: Tire Setup. We've moved away from static descriptors. Our new logic calculates suggested PSI/Bar by linking the Mud Risk and Surface Type to the bike's mechanical capabilities.
"We aren't asking the AI to remember the trail; we are giving it eyes on the ground."
Next Mission: Sentinel-2
The roadmap is clear. We are currently researching the integration of NASA Sentinel-2 multispectral imagery. This will allow the MCP server to "see" moisture levels directly from space via NDMI (Moisture Index), validating our weather-based predictions with orbital ground truth.
Ready to shred. The mission is just beginning.