Emotional trajectory

A single mood check-in is a snapshot. A string of them, read together with what you were actually saying at the time, starts to look more like a line — and lines are easier to act on than snapshots. Emotional trajectory is that line: a read on how your tone has been moving, session over session.

What it actually is

Two different signals feed this insight. The first is deliberate: the 1–5 mood rating you give during a check-in, which is exactly what it looks like — a number you chose. The second is inferred: the tone of what you actually wrote or said in your saved session summaries, read for warmth, tension, resignation, relief, and the like. Neither one is treated as more "true" than the other. They're shown together because they sometimes agree and sometimes don't, and the gap between them can be informative on its own — a 4/5 rating paired with a summary that reads as flat or tense is worth noticing.

How it's computed

The mood number is stored exactly as you gave it — no interpretation happening there. For tone, an AI model reads each session summary and estimates where it falls on a rough emotional register, then those estimates get plotted alongside your ratings across time. There's no hidden scoring against a clinical scale; it's a relative read, useful mainly in how it moves compared to your own baseline, not compared to anyone else's.

How to read it

Look at direction and duration more than any single point. A rough day that shows up as a dip is normal and expected — moods move. A trajectory that's been sliding gently for three or four weeks in a row is a different kind of pattern, and one worth naming out loud, to yourself or to someone else, rather than waiting for it to resolve on its own.

It's also fine, and common, for the inferred tone line and the mood rating line to diverge sometimes. People rate a hard day a 3 out of habit, or write about something difficult in a matter-of-fact way. Both lines are approximations of an internal state that only you actually have access to.

What it can't tell you

This isn't a mood disorder screening tool, and it isn't trying to be. It can't distinguish between a rough patch and something that needs clinical attention — that distinction depends on context this feature doesn't have, and on judgment it isn't equipped to make. It also only reflects the days you checked in or had a session; gaps in your usage show up as gaps in the line, not as neutral days. If the trend looks concerning to you, that's worth trusting over the chart either way — see mood patterns for how Therapy handles what looks like a sustained decline, and please don't wait on a graph before reaching out to someone if you need to.