See every session over a window

End of mesocycle. You don't want the summary — you want the list. Every session you ran, in order, with the headline numbers per session. The view you scan looking for the week that felt off, or the run of sessions where the volume climbed.

What the agent needs to hear

Behind your sentence, the agent works backwards to two pieces: the window and what to show per session. Window defaults to the last 3 months. Per-session columns default to date, name, total volume, duration, and PRs landed.

Window phrasings come in three shapes. Named ranges read intent (« since the routine activated »). Relative spans are the most common (« last 8 weeks »). Explicit dates pin the edges when you need precision.

Columns are flexible too. The default five answer most questions, but you can ask for more (« add average RPE », « show which exercises were skipped »). Name what you want to scan for and the column appears.

Window, per-session columns, and ordering

The window is the axis that does the most work. A short window is for recent texture — what just happened. A long window (« the whole mesocycle ») is for pattern — accumulation, drift, streaks. Pick the one that matches the question; the wrong window buries the answer.

Ordering is most recent first by default — you scan from today backwards, which is how memory works. Flip it with « oldest first » when you want the chronological story: where did the volume start, where did it land.

Per-session columns stay minimal by default so the list reads at a glance. Widening helps for specific reads (« add cardio sessions », « show readiness for each », « include cancelled and skipped »). The timeline is a list view, not a chart — the value is in scanning rows.

How the timeline differs from the weekly recap and the PR ledger

Three nearby reads, three different shapes. The weekly recap is one week compressed into one card — aggregated, dense, oriented to next Monday. The PR ledger is the all-time best per exercise — wide, sparse, the strength snapshot. The timeline is every session in order — long, scannable, pattern-focused.

Use the timeline when you’re looking for a specific session you half-remember (« the day the volume was huge »). Use it when you want to see streaks and gaps visually — a 6-week consistent run, then a 2-week hole. The timeline answers questions about rhythm, sequence, the shape of a block — questions the other two can’t.

When the agent gets it wrong

The list is where you spot the misses. Three patterns dominate. The window pulls in sessions you’d rather exclude — a deload block skewing the volume column, an injury week dragging the average. Carve it out: « exclude weeks 8 and 9 ».

Wrong columns is the second. You wanted RPE and the card shows duration. Name what’s missing and the agent re-pulls. The third is silent drops — cancelled sessions vanish when they’re part of the pattern you’re scanning for. Say « include cancelled ones — that’s part of what I’m looking at » and they reappear, tagged.

What makes the timeline worth scanning

Three things decide whether the timeline tells you something useful: the window covers the period the question lives in (« when did volume start climbing » needs the full mesocycle, not just the last month), the columns match the pattern you’re looking for (volume for accumulation, PRs for breakthroughs, readiness for behavior), and excluded sessions are surfaced as excluded, not silently dropped (skipped and cancelled days are part of the story when you’re scanning for gaps).

The timeline is the zoomed-out movie between week-level recaps and lift-level ledgers — the view where blocks become visible.