Love Meter: how much love is in your chat, with the numbers in front

You upload the exported chat and the engine converts each signal into a rate per thousand messages, which is compared against a fixed calibration curve. This means two important things: that a conversation of 300 messages and another of 80,000 yield comparable results, and that your score does not depend on what other couples have gotten. There is no user ranking.

The Love Score combines affection —explicit affection, romantic language, support, good morning and good night greetings, inquiries about each other— with intensity —volume, day-to-day continuity, response speed, nighttime chatting, and laughter— and falls into one of four levels: FRIENDS below 35, INTEREST up to 55, LOVE up to 75, and PASSION above. Below, the Relationship Profile scores five indicators separately: Affection, Passion, Fun, Communication, and Support.

Each indicator opens its “why”: the contributors with their actual counts and the exact definition used for measurement. Laughter, for example, is not counted just for appearing: its length is also measured, because “hahahaha” is not the same as “haha.” The response time is the median, not the average, and excludes waits of over 24 hours to prevent a night's sleep from distorting the result. The AI does not participate in the calculation: if you ask for a narrative, it writes the sentence above based on the already calculated figures.

The Danger Meter is calculated separately and never touches the Love Score: it measures jealousy, conflict, ignored messages, long silences, tone changes, and imbalance, with its own levels. And there is an honesty brake: below 300 messages or 30 days, aiWapp tells you that the reading is unreliable instead of making up a round number.