Strava is launching a new MCP Connector that gives subscribers a way to connect their Strava training history directly to Claude, Anthropic’s AI assistant.
Strava MCP Connector Strava announced the new MCP Connector on June 1, giving subscribers a way to ask natural-language questions about their own activity data through Claude. The feature is beginning to roll out globally and is included with Strava subscriptions.

That may sound technical, but the idea is fairly simple: instead of exporting your Strava data, uploading files, or manually copying workouts into an AI chatbot, the MCP Connector lets Claude securely read your Strava data and answer questions based on your actual training history.

What Is MCP?

MCP stands for Model Context Protocol. In plain English, it is a way for an AI assistant to connect to an outside app or data source in a structured way.

In this case, Strava is the data source and Claude is the AI assistant. Once connected, Claude can review your Strava activity history and use it to answer questions about your training.

That could include things like weekly mileage, pace trends, heart rate patterns, cycling power, training load, activity history, gear usage, and cross-training. Strava says the connector can access live Strava data rather than relying on a one-time static export.

What You Can Ask It

You can ask questions about their activities and health in normal language. For example, a runner could ask whether their easy runs are actually easy enough, how their mileage has changed in the lead-up to recent races, or whether their long runs are trending in the right direction.

A cyclist could ask how power has changed over time, whether a specific block of training has improved performance, or how different types of rides affect fitness.

Someone who mixes running, cycling, strength training, and other activities could ask how cross-training is affecting their overall progress.

Strava already has a lot of data, but most users do not have an easy way to interrogate that data beyond the charts and features Strava chooses to show inside the app. This gives subscribers a more flexible way to ask their own questions.

What Strava Data Can It Use?

Strava says the MCP Connector can provide access to detailed activity data, including per-second stream data such as heart rate and pace, GPS data for geographic analysis, cycling power data, and club and event data.

Strava’s support page also says the connector can currently be used for activity history, fitness trends, readiness, goal planning, cross-sport analysis, and gear.

How It Works

The Strava MCP Connector is available through Claude. Strava says users connect their accounts through OAuth, which is the same general authorization flow used by many third-party app connections. Connection instructions here.

Once connected, subscribers can access the Strava MCP Connector through Claude.ai, Claude’s desktop experience, or Claude Code. Strava says it is launching first with Anthropic’s Claude and is looking to support other AI clients in the future.

The connector is read-only, which means Claude can analyze Strava data but cannot upload, edit, or delete activities. Strava also says access can be revoked from Strava settings.

What This Means for Fitness Data

Fitness data is becoming easier to use outside of the app where it was originally collected.

Most fitness platforms already analyze your data in some way. Garmin Connect, Strava, Apple Fitness, Whoop, Oura, and others all have their own dashboards, charts, summaries, and insights (some of which leverage AI already). But those insights usually live inside each company’s own app.

By connecting Strava data to Claude, users may be able to put their training history in the context of other things they already use Claude for, including broader health questions, personal goals, race planning, lifestyle context, nutrition, recovery, scheduling, or other personal information they choose to share with the AI assistant.

For example, a runner could ask whether their mileage buildup looks aggressive given their recent sleep, stress, travel, or injury history. A cyclist could ask how their power trends compare with a specific training block. Someone training across multiple sports could ask how running, cycling, strength work, and recovery are fitting together.

The Bottom Line

Strava’s MCP Connector gives subscribers a new way to analyze their training history through Claude.

It is read-only, limited to Claude at launch, and only available to Strava subscribers. It also does not make Claude a perfect coach. AI-generated training advice can still be wrong, incomplete, or missing important context.

For Garmin users, the most interesting part may be that Garmin’s own analysis still lives in Garmin Connect (and requires a Connect+ subscription), while Strava’s MCP Connector could let athletes bring synced Garmin workout data into a broader AI conversation.

0
comments · 0 threads

Join the conversation. A free account lets you comment, reply and say how updates are running.

Sign up freeSign in

    No comments yet. Start the conversation.