ABOUT
What this is, and how the guides are checked.
An independent tool for the public side of YouTube — comments you can read, search, export, and captions you can download — plus a set of guides whose claims are checked against the product before they are published.
What this site is
A single place to work with the public side of YouTube: open a comment thread and read it in order, search inside it for a phrase, track down one comment by the name of whoever wrote it, take the thread out as CSV or JSON, and collect up to ten videos in one export. The same site reads a video's subtitles from the caption track YouTube already has and saves them as TXT, SRT or VTT. An optional AI reading summarises a thread and quotes the comments behind each claim.
Where the data comes from
Comments, replies and caption tracks are read through the official YouTube Data API. Nothing is scraped out of the watch page and nothing comes from a private account. That also sets the boundary: comments that were deleted, held for review, restricted to members, or posted on a video with comments turned off cannot be returned, and a large thread may not come back in one pass. Every result reports the coverage it actually reached rather than implying it saw everything.
How the guides are written
Each guide is written against the running product, not from memory. Claims about comment limits, credit costs, file formats, date handling and retention are checked before publishing, and when the product changes the guide is corrected in place and its updated date changes with it. A guide never describes a feature this site does not have — where something is not possible, the guide says so and explains the workaround instead of inventing a button.
What a credit buys
One credit opens a full comment thread, one credit reads a video's subtitles, and three credits buy the deeper AI reading of a thread. A new account starts with five. An extraction that fails is not charged. Credits are spent only after the work succeeds, and the exact prices are on the pricing page rather than buried in a plan table.
What this site does not do
It does not bypass privacy settings, moderation, deleted content or platform limits. It does not download video or audio, and it does not transcribe audio: subtitles come from caption tracks that already exist, which is why a video with no captions has nothing to download. It does not claim a thread is complete when YouTube returned only part of it.
How your data is handled
Comment bodies from a saved job are not stored. Export archives can be downloaded for 24 hours, after which the links expire on their own. A subtitle reading keeps the caption text you opened so the same reading can be reopened without spending another credit. The privacy page and the data policy describe the full set of what is kept and what is not.
This page was last checked on 21 November 2026 against the running product. If a guide and the tool ever disagree, the tool is right — and the guide is the thing that needs fixing.