It starts with one video, chosen by you. You pick a video where you have something real to say — a point you agree with, a question the video raises, a detail worth answering. Tarelivo works one video at a time because a comment worth reading is always about one video, never about a batch.
The draft is written for that video alone. It names something specific from it — a moment, a claim, a detail — and stays in the voice of a viewer talking to the creator, not a pitch. If the same words could sit under any video, the draft has not done its job yet; that is the first thing the rest of the process checks.
Then the draft is scored, not posted. The quality score reads the draft for specificity, relevance, tone, and shape, and flags links, self-promotion, and generic phrasing. A low score does not delete the draft — it points at the weakest part so you know what to fix before a person reads it.
Then a person reads it the way the creator would. Approval asks one question: could this only have been written by someone who watched this video? If yes, it can be approved; if no, it goes back for an edit or gets rejected. Nothing skips this step, and no score can approve a draft on its own.
An approved draft can be scheduled, but scheduling only sets the time. It does not change the words, and it does not revive a rejected draft. If the video's context changes before the scheduled time — a new pinned comment, a correction, a follow-up video — the approval is worth revisiting before anything goes out.
What the journey does not promise is as important as what it does. No step predicts replies, pins, subscribers, or views, and finishing the journey does not make a comment welcome — that is decided by the creator and their audience, every time. What it does is simpler: one specific comment, written for one video, checked twice, posted once by a person who read it. That is the whole job.