Log an AMRAP set (as many reps as possible)

Last set of the exercise. You loaded the weight and pushed until the bar wouldn't go up clean — twelve reps, the thirteenth wasn't going to happen. AMRAP isn't a target, it's an outcome. You log what came out.

What the agent needs to hear

Behind your sentence, the agent works backwards to three pieces: which exercise, the weight on the bar, and the rep count you finished with. The fourth piece — that this was max-effort — comes from the words AMRAP or « max » or « as many as possible ». You don’t have to spell every piece out: the active exercise gets assumed, so « AMRAP at 70 — 12 » is enough if you’re still on bench.

The agent shows you a card with the set tagged as AMRAP, the weight, and the rep count — and waits for a nod. One row, but the tag is what makes it read differently than the working set above it.

The rep count is the outcome

Regular working sets log what was planned. An AMRAP logs what came out. If the program asked for 8+ and you got 12, log 12 — the gap between planned and achieved is the point of the set. If you wanted 12 and stopped at 10 because the rep before was ugly, log 10. The agent does not coach you on whether you stopped too early; it logs the number you said.

There is no “target” field on an AMRAP. The weight is fixed, the rep count is variable, and the variability is the data. Pretending you hit a clean 12 when the eleventh was a grind is the kind of small inflation that quietly skews proximity-to-failure estimates over weeks.

What this signal is for

An AMRAP-tagged set is read differently than a regular set at the same weight. The system knows the rep count is a maximum, not the midpoint of a working range — and that’s what feeds the proximity-to-failure estimate, the rep PR ledger (« best AMRAP at 70 was 12 reps »), and intensity calibration over time. Hit 14 reps at 70 next week and the system reads progress against the previous 12, not against an averaged set count.

Logged as a regular set without the AMRAP tag, the same 12 reps at 70 looks like a normal working set with a high rep count — useful, but the max-effort signal is gone, and the progression story flattens.

When the agent gets it wrong

An AMRAP can’t be re-done, so the card is the moment to verify the rep count is real. If the count is off, say the right one: « it was 11, not 12 — the last one didn’t lock out ». If the agent missed the AMRAP intent and logged it as a regular set, push back: « that was AMRAP, not a regular set — tag it ». The card re-renders with the tag.

If the weight is off — you said 70 but the agent picked up 75 from the previous sentence — name it cleanly: « weight was 70, not 75 ». And if the agent treated the AMRAP as the start of a cluster (the word « drop » or « rest » appeared somewhere earlier in the chat), correct it: « one AMRAP set, that’s it ».

What makes the log worth keeping

Three things decide whether this AMRAP helps you later: the tag is preserved (so the system reads it as max-effort, not regular working), the rep count is the true outcome (not an aspirational number — what actually locked out), and the weight is what was loaded (so rep PR at that weight has a real anchor). The system uses these for proximity-to-failure, rep PRs, and intensity progression; a regular-set log of the same 12 reps loses the failure-anchored signal entirely. Say the set the way it ended, including the AMRAP word, and the rest follows.