Go Beyond a Single Ride Review

Connect Garmin or Intervals.icu, then use RideReviewer conversationally in ChatGPT or Claude to review a ride, compare similar workouts, test a pattern, and adapt what you do next.

Get 3 Free ReviewsEmail → access code → ChatGPT or Claude → ride source. No credit card required.

Move from a ride to a testable decision

Use the prompts in order or jump to the question you need now. Keep the evidence and its limits visible at every step.

  1. 01Rides
  2. 02Repeated evidence
  3. 03Likely limiter
  4. 04Uncertainty
  5. 05Next experiment

Start with one ride

Check whether the ride matched its purpose, identify the strongest available evidence, and choose one practical change to test.

Try this prompt
Review my latest ride. Start with what the data shows, separate observations from interpretation, and give me one practical change to test next.

Compare 2–5 similar rides

Compare rides with a similar goal. Arbitrary activities can make changes in terrain, duration, or workout purpose look like progress or decline.

Try this prompt
Compare my latest endurance ride with up to four similar recent rides. Check duration, power when available, heart-rate response, cadence and drift. Tell me what actually changed.

Find a repeated pattern

Ask for repeated evidence before accepting a limiter. Available recent rides can support a working hypothesis, not exact physiological causality.

Try this prompt
Across my available recent rides, find one pattern that repeats. Show which rides provide the evidence, suggest the most likely limiter, state the main uncertainty, and give me one experiment that could confirm or change the conclusion.

Decide the next 1–3 sessions

Use a short horizon that can adapt to new evidence instead of asking for a rigid long-term calendar.

Try this prompt
Based on my recent rides and current goal, suggest my next one to three sessions. Explain the purpose of each and what would make you change the recommendation.

Add goal and real-life context

Share only what matters: your event and date, weekly availability, terrain, session duration, power-meter and FTP or zones, fatigue, unusual sensations, sleep, fueling, weather, or time constraints when relevant.

Try this prompt
My goal is [goal or event] on [date]. I can train [days or hours] this week, usually on [terrain], and my sessions can last [duration]. I [do/do not] have power data; my known FTP or zones are [details]. Use this context with my available rides and tell me what matters most now.

Adapt after reality changes

Revise the next few sessions after a missed ride, fatigue, or changed availability while keeping the decision with you.

Try this prompt
I missed [session], my fatigue feels [description], and I now have [availability]. Reassess my next one to three sessions using the rides and context available here. Keep, replace, or remove each session and explain why.

Run a weekly reflection

Use retrievable rides to review what improved, what repeated, what is still uncertain, and what to test next week.

Try this prompt
Run a weekly reflection using the recent rides you can retrieve. What improved, what pattern repeated, what remains uncertain, and what single experiment should I test next week? Cite the rides that provide the strongest evidence.

Challenge the first answer

A useful coaching answer should survive questions about evidence, alternatives, missing data, and a more conservative interpretation.

Try this prompt
Challenge your first interpretation. Which rides provide the strongest evidence? What else could explain the pattern, what data is missing, and would the conclusion change without power data? Show the comparison in a table and give me a more conservative recommendation.

Use context from this conversation

RideReviewer can use the goal, constraints, sensations, and other relevant context you provide in the current conversation alongside connected ride data. It does not yet keep durable Athlete Memory across separate conversations, so restate important context when you start a new one.

Connect, switch, and refresh your ride source

Ask RideReviewer in ChatGPT or Claude to manage your integrations whenever you need the secure source-management page. It creates a fresh short-lived link; an expired or already used link cannot be reused, so ask again for a new one.

  1. 1

    Choose an active source

    Connect Garmin or authorize Intervals.icu with read-only access, then activate the source you want RideReviewer to use. Only one source is active at a time, and Garmin and Intervals.icu histories stay separate.

  2. 2

    Import and review

    Intervals.icu initially imports at most 30 completed cycling activities from a bounded 365-day search. Garmin follows its existing connected-history flow. Missing or unavailable metrics remain unavailable.

  3. 3

    Refresh after later rides

    Ask for a manual refresh when you want newer Intervals.icu rides imported. RideReviewer does not poll Intervals.icu or refresh it automatically in the background.

    Source management remains available after your three free reviews are used. Trial exhaustion limits new Ride Reviews; it does not prevent you from reconnecting, activating, refreshing, or disconnecting a ride source.

Capabilities and limits

RideReviewer works with the recent rides and metrics available through the current client and your active Garmin or Intervals.icu connection. Missing or uncertain data stays missing or uncertain.

Supported today

  • Review one ride
  • Compare 2–5 rides
  • Identify repeated evidence and a likely limiter from available recent rides
  • Recommend or adapt the next 1–3 sessions
  • Use goals and context supplied in the current conversation
  • Explain uncertainty and missing data

Do not claim

  • Medical diagnosis or injury advice
  • Exact physiological causality or guaranteed improvement
  • Automatically available sleep, nutrition, sensations, or weather
  • Durable cross-conversation Athlete Memory
  • Autonomous scheduled tasks or automatic background ride analysis
  • A rigid or perfect long-term training plan