TikTok Comments Scraper

It turns a video's comment section into rows instead of a scroll: the text, who wrote it, what it earned, and where it sits in the thread. Tell AllyHub which videos in a sentence or two, and it works through several of them, or a creator's whole library, opening the reply threads as it goes.

Yes. A single video with the standard comment fields runs free. What the paid plans carry is the bigger shape of the work: many videos in one job, deeper reply expansion, filtering at fetch, scheduled re-pulls, and the analysis pass on top.

No tool can promise that, and the ones implying otherwise are not describing how TikTok behaves. The section loads as you scroll, replies sit behind a link, and the order shifts between loads. AllyHub goes as far as the page will go and then tells you the number it returned and when it ran, so what you hold is a sample with a date on it rather than an unqualified total.

Yes, and that is the usual shape of the job. Give it a handle, a list of links, or the videos where a term shows up, and every row comes back tagged with the video it belongs to, so you can read one section or compare across all of them.

No. TikTok's official comments endpoint is part of a research program with an eligibility bar most people cannot clear, and it is not what this uses. AllyHub reads the public page the way a person would, so there is no developer account, no application, and nothing to write.

It reads only what a visitor sees without signing in, and it does not go behind a login or build profiles of the people commenting. Comments are still written by real people, so treat the output as personal data: keep what the research question needs, and check your own obligations before you store or publish it.

Scrolling gives you an impression and no way to show your work. This gives you the comments, their counts, their dates, and the repeats grouped together, kept as a Playbook that compounds with every task — so the next video's section is one sentence away instead of another hour.

That is the same job under another name, and for most people the export is the whole point of the pull. One row per comment, carrying the columns you asked for rather than every field it holds — the system waiting downstream usually has opinions about that. It is written as UTF-8 with separators escaped, so a comma inside a comment stays in its own cell and emoji survive the trip instead of arriving as boxes, and posting times come out as real dates rather than ten-digit numbers. When several videos go into one job, rows are deduplicated by comment, so a thread quoted under two of them does not land twice.