Most runners have accumulated far more training data than they will ever properly look at.
Runs, heart rate, power, cadence, race performances, training plans and months of workout history can all sit inside an app, quietly gathering dust once the next session begins.
Stryd now wants artificial intelligence to make sense of it.
The running power company launched the Stryd MCP Open Beta on November 3rd, allowing runners to connect their Stryd account directly to compatible AI assistants including ChatGPT and Claude.
Once connected, runners can simply ask questions about their training in normal language.
Instead of manually comparing several months of workouts, somebody could ask where their power began falling during a marathon, how that compared with their long runs or how their current fitness compares with the beginning of their training block.
The AI can then go looking through the underlying Stryd data and build its answer from the runner's actual training history.
What Can It Actually See?
Quite a lot.
Stryd says a connected assistant can access recorded activities, planned workouts, Critical Power history, races on the runner's calendar, training plans and race prediction information.
It can also see equipment information, including the distance recorded against individual pairs of shoes.
Go deeper into an individual activity and the assistant can access lap information including pace, power, heart rate, cadence, ground contact time and leg spring stiffness.
It can even examine second by second activity data when more detailed analysis is required.
That creates some fairly interesting possibilities.
Imagine finishing the Dublin Marathon and asking:
At what point did my power begin to fall significantly?
You could then follow that with:
Was the same pattern present during my longest training runs?
Or:
How did my heart rate and power during the final 10 kilometres compare with the first 30 kilometres?
Instead of opening several activities and trying to piece the answer together yourself, the AI can pull information from the relevant sessions and compare them.
You Could Ask Why A Session Went Wrong
The feature may actually prove more useful during ordinary training than after a major race.
A runner completing an interval session could ask how consistent their repetitions were.
The assistant can inspect each lap and compare time, pace, average power, maximum power, heart rate and cadence.
So somebody struggling through the final few repetitions could ask:
Did my power actually drop during the final three reps, or did they only feel harder because my heart rate was rising?
That is a fundamentally different way of interacting with running data.
Most training platforms present information visually through graphs, numbers and charts. The runner still has to interpret what those numbers mean.
Stryd is now allowing the runner to start with the question instead.
It Can Look Across Months Of Training
Perhaps the more powerful use comes when the assistant is given a much wider period to examine.
Stryd suggests asking questions about trends across an entire six month training block, changes in running power or which training sessions most closely resembled a later race performance.
That could make historical training information considerably more useful.
A marathon runner might ask how their Critical Power has changed since beginning a programme.
Another could compare their latest long run with one completed on the same route several months earlier.
Someone returning to racing could ask how their current training volume compares with the period preceding a previous personal best.
The data already existed.
What has changed is how easily somebody can interrogate it.
It Can Even Give Race Power Predictions
Stryd's system also gives compatible assistants access to its race prediction tools.
An assistant can request a predicted finishing time and recommended target power using the runner's current Critical Power and Stryd's underlying calculations.
That means somebody preparing for a half marathon could ask what power they should target based on their current data rather than going into Stryd and finding the calculation themselves.
The assistant can then use other available information to provide context around that number.
For runners already accustomed to racing with power, that could make planning considerably quicker.
Your Training Plan Is Available Too
The connection is not limited to completed activities.
AI assistants can also read workouts scheduled on a runner's calendar and information from their current Stryd training plan.
That means questions can be forward looking too.
You could ask what sessions are coming this week, what stage of the training plan you have reached or how the next few workouts fit with the work already completed.
Stryd says the assistant can see the goal and dates of the current plan, its present phase and the runner's progress through it.
There is an obvious appeal here for somebody who has spent months collecting data but rarely ventures beyond checking the summary after each run.
It Does Not Get Control Of Your Account
There is an important distinction.
At present, the Stryd MCP beta is read only.
The AI can inspect information but cannot upload a workout, change settings, edit activities or delete anything from the account.
Stryd says the ability to write information back into the platform is planned for the future, but that will require the user to turn the functionality on themselves.
Access is also limited to the runner's own information, apart from athletes who have deliberately shared their profile with that person through Stryd's coaching system.
Connecting the service requires the user to sign into Stryd and approve access.
There Are A Few Requirements
The new feature is currently part of the Stryd Membership, so simply owning a Stryd foot pod is not enough.
It is also an open beta.
Stryd warns that available tools, data fields and usage limits could change as development continues. Access can also be limited or temporarily withdrawn during the beta period.
For ChatGPT users, Stryd's instructions require a plan that supports custom connectors.
Claude, Codex, Cursor and other compatible systems can also connect to Stryd's service.
The Interesting Part Is Not Really The AI
There is no shortage of products currently adding the letters AI to their marketing.
What makes the Stryd implementation more interesting is the data sitting underneath it.
A generic AI assistant knows very little about how your Sunday long run actually went.
Connect it to years of genuine Stryd activities and suddenly it can inspect your pace, power, heart rate, cadence, Critical Power history and individual laps before answering.
That does not mean every conclusion it produces should be accepted without question.
Training data can be messy, sensors can produce poor readings and an AI generated interpretation is not a substitute for an experienced coach who understands the athlete behind the numbers.
But it could make the enormous volume of information runners already collect far easier to use.
For years, running technology has concentrated on gathering more data.
Stryd's latest experiment asks a rather different question.
Now that we have all this information, can we finally make it easier to understand?
