STAS knowledge base

Detailed workout breakdown

A detailed workout breakdown takes a close look at one specific session. It is not the compact preview you see in a summary, and it is not a list of workouts over a date range. It opens the workout interval by interval, so you can understand how it was actually executed.

A workout list gives you context across many sessions. A detailed breakdown zooms in on one workout: every work interval, every recovery segment and, when needed, second-by-second data streams.

What it is

STAS can show training data at several levels. This page explains the deepest one.

Three different ways to request data

These levels are not separate tabs in the interface. They are different ways ChatGPT or Claude can ask STAS for context.

The summary gives the AI a compact preview of several recent workouts. A workout request over a date range returns many sessions at once, with metrics, zones, splits and reports. A detailed workout request opens one session interval by interval.

How this differs from a workout list

A workout list over a date range and a detailed breakdown of one workout are two different STAS requests. They answer different questions.

Workouts over a date range

The list gives an overview of many workouts at once: how much load you accumulated, how trends changed, where the plan differed from what happened, and how volume and intensity were distributed across a week or month.

One detailed workout

A detailed breakdown opens one workout by intervals: each work segment and each recovery segment separately, with its own duration, pace, heart rate or power.

What the breakdown includes

The breakdown uses everything available in the workout file and adds the full interval structure.

Main blocks

When the data was recorded, a detailed breakdown may include:

intervals — the full set of work and recovery segments, each with its own duration, pace, heart rate or power. This is the part that is not available in the date-range workout list;
splits — progression by kilometres or miles;
heart-rate zones — time in zones and threshold values;
general metrics — distance, duration, pace, average and maximum heart rate, training load and elevation gain;
running metrics — normalized pace, cadence and vertical oscillation;
cycling metrics — average and normalized power, FTP, intensity, TSS and power zones;
external assessment — a self-rating recorded by the device or app, such as RPE or perceived feel;
your workout report — your own notes on how the session actually went.

Deeper: data streams

For the deepest analysis, STAS can also request continuous data streams from Intervals.icu.

Continuous streams

In addition to the saved workout record, STAS can request heart rate, pace, power, cadence and elevation at each point of the workout, plus automatic lap detection.

This allows the AI to inspect execution down to seconds and metres.

How the AI gets the breakdown

The AI first finds the workout in a list, then requests the detailed breakdown for that specific session.

List first, detail second

To analyse a workout in detail, the AI first requests workouts over a date range, finds the right session, takes its identifier and then makes a separate request for the detailed breakdown.

“Break down my latest workout in detail.”

For a normal detailed review, the full workout record with intervals and metrics is usually enough. If second-by-second data is needed — heart rate, pace or power at each point — the AI can load those streams from Intervals.icu as well.

When to use a detailed breakdown

Not every question needs this level of detail. Sometimes a workout list over a date range is the better tool.

When one session matters

A detailed breakdown is useful when one workout matters: what exactly happened during a key session, how well you held the target segments, why the last set became harder, or where the watch may have split the workout structure incorrectly.

For questions about a whole period — total load, volume growth or intensity distribution — the date-range workout list is enough. Pulling one detailed workout into that context usually adds noise.

Limitations

A detailed breakdown can make the analysis more precise, but it is limited by what was actually recorded.

Important boundaries

The available data depends on the device and sport. If power was not recorded, there will be no power fields.
STAS does not invent metrics. If something was not recorded, it will not appear in the breakdown.
External assessment, such as RPE or perceived feel from a device or app, is not a fact and is not your own report. It can be considered, but it does not replace your notes.
Continuous data streams are not always available. This depends on the device and on Intervals.icu.
The breakdown is context for analysis, not a ready-made decision. The final decision is still yours.

The breakdown is about one workout

Detailed data is loaded for one specific session. For the bigger picture across a period, use the workout list over a date range. That keeps the context accurate and avoids overloading it.

FAQ

Short answers to common questions about detailed workout breakdowns.

How is this different from a workout list over a date range?

A date-range workout list gives an overview of many workouts: training load, trends, plan versus actual, and the distribution of volume and intensity. Each workout includes metrics, zones, splits and a short preview of the interval structure.

Are power, cadence and zone data always available?

No. The data depends on the device and sport.

What is an external assessment?

It is a self-rating recorded by a device or app, such as perceived exertion, RPE or overall feel.

Why use data streams?

They provide the deepest level of detail: heart rate, pace, power, cadence and elevation at each point of the workout, plus automatic lap detection.

Does the AI see detailed data for every workout at once?

No. The AI first sees a list of workouts over a date range. It loads the detailed breakdown only for a specific session, when a deeper look at one workout is useful.

Can I trust the interval structure from the watch?

Not always. Watches can split intervals incorrectly or lose part of the structure.