To track anything in ChatRPE, name the metric, say the value, and give the unit or scale: "headache 3 out of 10," "caffeine cutoff 2 p.m.," "grip 52 kg." There is no form to design first. The first entry sets the shape, later entries follow it, and the metric can be charted against your training, weight, or sleep.
Choose what you would use the record to decide before collecting it. A stable name and measurement method make later comparisons more useful than a growing collection of unrelated numbers.
Start with the decision, not the data
Tracking without a question is homework. Before adding a metric, finish this sentence: "If this changes, I will…"
If no result would change your behavior, skip it. A simple metric you log consistently beats a detailed schema you abandon.
Daily water is the plainest example of a metric that passes that test, since the decision it feeds is whether to carry a bottle tomorrow. The guide on how much water you should drink a day works through the target itself, the one-session weigh-in that tells you what you sweat, and the reminder and dashboard card that keep the number in front of you.
Let ChatRPE shape the fields
Most useful custom records need a date, value, short note, and one or two pieces of context. You do not have to design that up front. Describe what you want to watch and let the structure fall out of the description.
"Track my left-knee status after leg days. Use pain from 0–10, swelling yes/no, triggering exercise, and whether it changed the workout. Keep this structure for future entries."
Use one stable name for one idea. If knee_status becomes knee pain next week, ask ChatRPE to standardize it before analysis. Two names for one metric is the single most common reason a custom log stops being chartable.
It remembers the structure you invented
This is the part a "custom field" in a normal tracker cannot do. ChatRPE keeps a live index of the data types you have actually saved and the field paths inside each one (that knee_status has pain as a number, swelling as a boolean, exercise as text), and that index is loaded into the coach's context before the conversation starts. It is built from a recent sample of your entries, so a metric you have not touched in a very long time may need to be named once to bring it back into view, but for anything you are actively logging the coach already knows the shape.
Reuse the same metric name, units, and scale across entries. For example, "knee discomfort 3/10 after squats" should use the same field and definition each time. Check the saved structure if a later entry looks different.
Ask for the number, not just the chart
Once a custom metric has a few weeks behind it, ChatRPE can run real analysis on it rather than just displaying it: the average, total, minimum and maximum, the most recent value, how many entries went into it, and whether the direction over the window is increasing, decreasing, or flat, with a rough percent change. That works on any numeric field in any data type you invented, not just the built-in ones.
"What's my average knee pain score over the last month, and is it trending down since I changed my squat depth?"
Name the data type and the period you care about and the analysis answers that question precisely, which is how to interrogate a long-running metric one phase at a time. The direction it reports comes from comparing the earlier and later halves of that window, so treat it as the signal that tells you where to look closer.
Two details decide what those numbers mean. The average is taken per entry rather than per day, so a metric you record twice a day returns an average reading rather than an average day. And the summary reads at most your 100 most recent entries of that type, which is a month at three a day, so a longer window can quietly be answered from a shorter one. Ask which dates it covered whenever the window is the point.
A chart groups your entries into one point per day, and what happens inside a day depends on what the field is called. Names containing cal, energy, protein, fat, carb, or fiber are added together; every other field is averaged unless you ask for the total. Three coffees logged separately plot as your average cup rather than the day's caffeine, so say which one you want, or keep that metric to a single entry a day.
Run a real experiment
"Track caffeine cutoff, sleep quality from 1–5, and morning energy for 14 days" is better than "track sleep." It has a window, a small schema, and a decision attached.
ChatRPE can plot two saved metrics together, use separate axes for different scales, or draw a scatter plot. The sleep tracker guide shows the sleep-and-training example. Treat associations as questions to investigate rather than evidence that one measure caused the other.
Review noisy signals in batches. Then use small dashboards to keep only the views that help you decide.
Metrics with no number at all
Plenty of things worth tracking have no value to plot. You took the supplement, you did the mobility work, the headache happened. Those still log fine, and ChatRPE can chart them by frequency, counting entries per day, which is what turns a yes/no metric into a calendar heatmap of how often it actually occurred. That is the same machinery behind consistency tracking in the habit tracker guide, pointed at whatever you decided to watch.
Free-text belongs in the record too. A custom entry can carry a note, and notes are searchable by keyword later, so "what was I writing about my elbow in March" is a real question rather than a scroll. The workout journal guide covers the note side of the logbook in depth.
Your custom data is still yours
Anything you invent is included when you export. The download is a zip containing a CSV of your entries with their data types and fields, a Markdown file of your notes, and an index, so a metric you made up in week one comes out in the same file as your workouts and weigh-ins. See the data export guide for the full format and how often you can generate one.
Take the record to the appointment
A clinician owns the diagnosis, the condition, and any treatment decision that follows from it. What you can bring them is evidence: eleven weeks of a symptom score with dates on it, the sessions that preceded each flare, and the note you wrote the morning it was worst. Most appointments start with somebody trying to remember when a problem began, and yours does not have to.
Ask ChatRPE for the summary before you go, then chart the metric and download that chart as a CSV to hand over. An export works too: ask for that metric and period as a spreadsheet and you get a CSV of the logged entries themselves. The data export guide covers both routes and which file suits which reader.
Frequently asked questions
Can ChatRPE track data that is not fitness-specific?
Yes. Habits, symptoms, notes, expenses, tasks, and custom scores can all become structured entries when they support your goals or routine. The category name is free-form, so the metric does not have to resemble anything a fitness app ships with.
What is the best custom fitness tracker app?
The useful test is not whether an app lets you name a metric, which most do, but whether it can analyze the metric afterward and read it against everything else you log. ChatRPE keeps custom metrics in the same logbook as workouts, food, weight, and sleep, remembers the structure you defined, and can average, trend, and chart any numeric field you invented. That is what makes a custom metric worth logging rather than just worth naming.
Is there an app where I can track anything I want?
That is what this workflow is for. Describe the thing you want to watch in a sentence, agree on a simple structure, and log it in plain language from then on. There are no fixed categories to choose from and no schema to design in advance.
Do I need to define every field up front?
No. Start small and refine the structure after you see what is actually useful. Adding a field later is a sentence, and older entries simply do not carry it.
How do I avoid tracking too much?
Use one rule: if no trend in the metric would change your behavior, it is noise. Do not track it. Two or three custom metrics with a decision attached beat ten you stop filling in by week three, and the daily coaching workflow is a better home for anything you only want to mention in passing.
Can I compare a custom metric against my training or weight?
Yes, and that is usually the reason to track it. Because everything lives in one logbook, you can ask for a custom metric and a standard one on the same chart (pain scores against training volume, a supplement against sleep quality, stress against daily calories) with separate axes or a scatter plot depending on what you are looking for.