A summary, not a built-in AI coach.
Strength Direct generates the summary on your device. It turns your workout history and chosen training goal into text you can review, copy, or share with an external AI assistant. It does not upload it to an AI service or require an AI account.
You can make useful training decisions with a clear program and consistent logging alone. An AI conversation is another perspective, not a replacement for that foundation.
Choose the question before sharing the data.
- Log some workouts first. A useful review needs completed training data.
- Open Analytics → AI Recommendations. Tap the blue sparkle button to generate a preview.
- Open Settings in the summary. Choose the time window you want to review and the training goal you want feedback on. Either change regenerates the text.
- Review, then Copy or Share. Add context your assistant cannot know, such as schedule, equipment, recovery, and constraints.
- The time window determines which training period you want to discuss.
- The training goal gives the conversation a purpose rather than inviting generic advice.
If your selected window contains no workouts, the export uses up to your ten latest older workouts instead and says so explicitly. Check the dates before asking about your current routine.
Read the preview before you send it.
- Inspect the text on your device before using Copy or Share.
- Body weight, workout names, and notes can be personal. Remove anything you do not want to send.
You can copy the text into an editor and remove anything you do not want to send. An external assistant is a separate service with its own account requirements and privacy policy.
Read the privacy policyAsk for evidence and one useful next step.
The summary’s opening prompt asks the assistant to cite your log, separate observations from hypotheses, preserve what works, and ask questions when context is missing. It asks for up to three changes with progression rules for the next 2–4 weeks.
“Which comparable sessions show a change in reps or load? What evidence supports your suggestion, what is still unknown, and what one change should I try within my program?”
Missing RIR means unknown, not failure. Unlogged training, recovery, and adherence are also unknown. Ask the assistant to explain its evidence rather than treating gaps as facts.
AI suggestions can be wrong. Use them as a starting point for discussion, not a diagnosis or a prescription. Consult an appropriate professional for pain, injury, or medical concerns.
Go deeper: what the summary includes
- Training context: goal, current profile body weight, exported dates, workout count, and average frequency.
- Weekly volume by exercise group: average volume, sets per week, and the split between 1–6, 7–12, and 13+ reps.
- Exercise progression: all-time best estimated 1RM, window trend when enough data exists, dated first and latest estimates, session count, and latest sets.
- Timed training, when logged: weekly duration and distance, plus best set duration or session-average speed.
- Workout details: newest-first workouts, completed working and drop sets, supersets, and workout and exercise notes. RIR is included when enabled.
Warm-up and incomplete sets are excluded; identical consecutive sets are compressed. The export uses your app language and selected units. Your own names and notes stay as written.
Go deeper: load conventions and trend limits
Dumbbell loads are per dumbbell. Bodyweight sets show extra weight added to the workout’s recorded body weight; their e1RM uses body weight plus extra weight. In the app, e1RM is displayed as total or extra weight depending on the exercise type; the AI export reports bodyweight e1RM as total load.
Weekly averages cover the whole exported period, not necessarily the current routine. Exercise-group totals are not muscle-specific.
An e1RM is an estimate, not a tested maximum or an effort-adjusted result. A window trend requires usable estimates on at least four distinct dates spanning at least 28 days; otherwise it reports insufficient data. It describes logged performance, not proven strength change, and can hide reversals.
