Hello Strava Developer Community and Strava team,
I’m seeking clarification before inviting any additional athletes to a private pilot.
The application is a direct Strava API integration that creates a private historical Runner Profile for the same athlete who authorized access. It does not display or disclose one athlete’s data to another user.
The Runner Profile uses deterministic, versioned rules to produce descriptive analyses of:
-
training history and data coverage;
-
consistency and training load over fixed periods;
-
personal running progression;
-
historical easy-run patterns;
-
personal intensity distribution when sufficient heart-rate evidence is available.
The implementation does not use an LLM, generative AI, machine-learning model, embeddings, vector store or RAG. Neither raw nor derived Strava data is sent to an AI provider or used for AI/ML training, evaluation, grounding or model improvement.
Privacy and security boundaries:
-
Every athlete can access only their own data and report.
-
OAuth credentials are encrypted.
-
Routes, GPS coordinates and polylines are not requested or stored.
-
Data is not sold, used for advertising or shared with third parties.
-
Disconnect and deauthorization delete the athlete’s stored data and derived reports.
-
Activity deletion and privacy-change webhooks remove affected data and invalidate derived reports.
The intended product may charge the athlete for the standalone personal coaching analysis, not for access to Strava or for functionality that reproduces the Strava service.
I am unsure how the following parts of the API Policy apply to this use case:
-
Does Section 5.4 permit a deterministic personal coaching analysis that is generated solely for and displayed solely to the same authenticated athlete?
-
Does Section 6.2 permit historical activity summaries and derived personal reports to be retained for longer than seven days, solely to provide the athlete’s ongoing Runner Profile and only until disconnect or deletion? If not, what retention design would Strava consider compliant for a historical coaching product?
-
Is charging for this standalone deterministic coaching analysis permitted under Section 5.8?
-
Since Developer Support redirected this question to the Community Hub, is there a formal channel through which we can obtain written permitted-use clarification before launching a small pilot?
I want to ensure the product is designed on the correct side of the current API Agreement and API Policy before proceeding. Guidance from a Strava staff member would be especially appreciated.
Thank you.
