Seeing all the comments — and knowing when “all” is really all.

“See all comments” is the most reasonable request anyone makes of a comment section, and the one the watch page is least built for. The page is a reader, not an archive — and the difference shows up the moment a thread matters.

What the page shows you

Under every video, the comment section starts small: a first batch of comments, loaded with the page. Everything after that arrives only when you ask — press Load more and another batch comes down. Switch the sort and the section starts over, discarding what was loaded. Replies stay folded under their parent comment until opened, one thread at a time.

This is a reading design, and for reading a conversation it works. The trouble is that it is also the only interface most people ever see, so “the comments” quietly becomes “the first few screens of the comments” — and nobody notices the substitution.

Why scrolling doesn’t finish the job

The obvious fix — keep pressing Load more until the section stops growing — works on small threads and quietly fails on big ones. You cannot tell the bottom from a pause. A sort switch erases your progress. And on threads with tens of thousands of comments, the page will not serve you the whole list however patient you are. The loop feels like progress and ends without a signal either way.

Reading the full thread outside the page

The way out is not to scroll harder but to read somewhere else — a reader that fetches the thread itself, and a file for what you want to keep. If what you are after is specifically the export — formats, fields, what ends up in each column — that has its own page; this one is about the reading side.

1. Load the thread in a tool that fetches all of itA reader built for whole threads pulls the public comments in its own batches and lays them out in one view — no Load more loop, no reset when you change the order. Searching inside the thread becomes possible once everything is on one screen.What to do: Open the video in the viewer and let it load: the whole public thread in one list, sortable five ways and searchable by phrase.
2. Export it as a file you keepReading is one thing; keeping is another. An export turns the thread into CSV or JSON with the same fields on every row — the version that survives sorting in a sheet, charting in a script, and the thread itself changing underneath you.What to do: Export as CSV for a spreadsheet pass or JSON for a script. The five columns are the same in both formats.
3. Check the coverage before calling it completeOn ordinary threads the export is the whole public thread. On giant ones, even a tool receives what the platform serves, batch by batch — so a big file is a large sample. Knowing which one you are holding is part of reading it honestly.What to do: Compare the export's row count against the video's comment counter and the tool's own coverage figure before treating the set as complete.

When “all” still isn’t everything

Every tool on earth reads the same public data, which means every tool shares the same blind spots. Comments held for review, deleted, filtered by the channel, or restricted to members never enter the public thread — they are not in the export because they are not served. The mechanics of those ceilings are collected on the comment-limits page.

Reading it end to end

One load in a reader gets you the whole thread; one export keeps it. Between the two, “see all comments” stops being a promise the page cannot keep and becomes a thing you did.

• Read it end to end — The full public thread in one view — loaded once, searchable, sortable five ways.
• Take it with you — CSV or JSON of everything reachable — the copy the thread cannot change.
• More than one video — When the thread is a channel — a whole back catalogue in one job.