Hi Strava team,
I'm building a personal coaching tool and want to make sure I'm reading the 2026 API terms correctly before relying on them. I'd really appreciate an official clarification on one specific, narrow use case.
The setup: a single athlete authorizes their coach to review the athlete's own Strava data on the athlete's behalf (via the athlete's own OAuth authorization). The coach stores that data and uses their own AI tools to analyze it, solely to give that athlete coaching feedback.
To be clear about what this is not: the data is never sold, never aggregated with other athletes, never exposed to any other user, and never used to train, fine-tune, or improve any model. It's one athlete's own data, used to coach that one athlete.
Where I'm stuck: reading the current API Policy, several clauses appear to prohibit this even in the single-athlete case — Section 5.3 (no use of Strava Data "in connection with the operation of any AI Application," including ingestion into a context window), Section 5.4 (no analytics / customer-insight generation; no combining Strava Data with other data), and Section 5.5 (no persistent storage beyond a 7-day cache). On a literal reading, a coach using AI to analyze their own athlete's data looks disallowed.
That seems practically inconsistent with the community Strava has said it wants to support. The announcement explicitly preserved "coaching platforms focused on providing feedback to users," and the new MCP is positioned for personal AI analysis of one's own data. A coach helping a single, consenting athlete with that athlete's own data feels squarely within both the spirit of "coaching" and of "personal analysis" — yet the letter of the policy appears to exclude it.
Could you clarify:
- Is a coach using AI to analyze a single consenting athlete's own data — not sold, not aggregated, not used for model training, shown only back to that athlete — permitted under the current terms?
- If not via the developer API, is there a sanctioned path for this single-athlete coaching use case — e.g., the athlete's own data export, an Extended Access arrangement, or the Strava MCP?
I genuinely want to build on the right side of the policy, so any official guidance would be hugely appreciated. Thank you.
