CyclingPlus – AI-Powered Cycling Insights
🚴 Inspiration
As a dedicated cycling enthusiast, I always hoped to get personalized feedback from real specialists. But deep, expert-level insights were always out of reach — until I turned to AI. That’s how CyclingPlus was born: an AI-driven companion that helps me understand my performance on a level no human coach could provide 24/7.
⸻
⚡ What It Does
CyclingPlus syncs your workouts directly from Strava and turns raw cycling logs into actionable insights using AI. It helps you analyze: • Power & cadence patterns • Fatigue indicators • Acceleration/deceleration trends • Strengths/weaknesses in your rides • Customized improvement suggestions
Think of it as having an expert cycling coach inside your Mac.
⸻
🛠 How We Built It • Kiro as the processing foundation • Codex and other LLMs for refining logic and enhancing analysis • Fully developed within the Apple ecosystem, optimized for macOS at this stage • A hybrid architecture combining Strava APIs + local analytics + AI reasoning
⸻
🧩 Challenges We Ran Into • Kiro couldn’t fully resolve Strava OAuth — requiring custom patching • Anthropic models performed poorly for structured sports analytics • Handling large workout datasets efficiently • Cross-model consistency for AI insights
⸻
🏆 Accomplishments We’re Proud Of • End-to-end platform engineering across the Apple stack • Fully working Strava → AI insights pipeline • Clean UI with fast data syncing • A real athlete-level analysis experience built by a solo developer
⸻
📚 What We Learned • AI alone isn’t enough — you need the right combination of tools • Collaboration between frameworks (Kiro, Codex, custom scripts) is essential • Cycling data is messy; building stable analytics takes engineering discipline • Testing with real ride logs matters far more than synthetic data
⸻
🚀 What’s Next for the iOS App
Upcoming features: • Enhanced data visualization (heatmaps, cadence-power scatter, elevation flow) • Acceleration & deceleration analysis for technical riding segments • Better climb/descend breakdowns • AI-powered training suggestions • Expansion to iOS + iPadOS versions
Log in or sign up for Devpost to join the conversation.