Running MCP
A personal tool idea connecting running data, Strava, and an assistant that can help me understand my training.
- My role
- Concept builder, API explorer, and runner
- When
- Current exploration
- Tools and themes
- MCP, Strava API, AI, Running data
Why it exists
Running creates a lot of personal data. Distance, pace, effort, consistency, and routes accumulate over time, but the questions I care about do not always fit a standard dashboard.
I wanted to explore a more conversational tool: something that could work with my running history and help me investigate patterns without turning every question into a spreadsheet.
What I worked on
I explored how Strava data, an AI assistant, and MCP could fit together as a personal tool. The work is part API exploration and part product question: what context should the assistant have, and which questions are worth making easier?
My perspective as a runner helps keep the idea specific. The goal is not a generic AI chat box beside a chart. It is a tool built around the small decisions and curiosities that appear during training.
Decisions that mattered
The assistant should stay grounded in actual activity data and make its reasoning legible. It should help compare, summarize, and notice, while leaving the runner in control of interpretation.
Keeping the experiment narrow also matters. A useful first version needs a small set of good questions more than a long list of AI features.
Where it stands
Running MCP is an active experiment. It represents how I like to learn emerging technology: connect it to a personal problem, build a focused version, and see where the idea becomes genuinely useful.
What I learned
Personal tools are a strong test for product honesty. If an interaction does not help with a real question I have, adding AI does not make it valuable.