Download YouTube Comments with Replies and Parent IDs

1. Collect comments and replies from a video

Paste a supported public video URL into TubeVox, sign in, review the extraction settings and start the job. At least 500 available credits are required to begin. Top-level comments and saved replies each cost 1 credit and count toward the extraction limit. Unused reserved credits return when processing settles.

Wait for the extraction to finish before preparing a final analysis. A saved dataset is limited by the extraction budget and the comments available through the source. Disabled comments, removed replies or restricted videos can prevent a conversation from being retrieved. Do not assume an export contains every comment ever posted.

2. Keep the identifiers in every format

Open the export dialog and choose Excel, CSV or JSON. Use all fields, or select at least Comment ID, Parent Comment ID, Comment and Type. Add Published and Author Channel ID if your analysis needs time or author information. Selecting only the comment text removes the information needed to join threads reliably.

Comment ID identifies the exported comment itself. Parent Comment ID identifies the top-level comment to which a reply belongs. A top-level comment has no parent value. These are YouTube identifiers; display names are not suitable substitutes because different people may share a name or change it.

3. Connect replies to their parent comments

Consider an illustrative dataset: a top-level row has Comment ID “sample-parent-1” and an empty parent. Two reply rows have IDs “sample-reply-1” and “sample-reply-2”, with Parent Comment ID “sample-parent-1”. Both replies belong to that same thread. These sample IDs explain the relationship and are not real YouTube comments.

In a spreadsheet, match each non-empty Parent Comment ID to the Comment ID column to retrieve the parent text. An exact-match lookup is appropriate. In a script, build a lookup keyed by Comment ID and resolve each reply’s Parent Comment ID against it. Keep IDs as text and keep the original columns alongside any derived thread labels.

A parent ID identifies the thread’s top-level comment. It does not prove that the reply directly addresses a particular preceding reply. Avoid inferring reply-to-reply relationships from row order, matching author names or mentions alone.

4. Understand missing parents and filtered exports

A replies-only export can contain valid parent IDs without the parent rows. Filtering by keyword, date or another condition can also remove the parent while retaining a matching reply. First export all saved comments with IDs and search that complete file for the referenced parent.

If the parent is absent from the full saved extraction, record the thread as incomplete. The extraction limit or source availability may explain the gap; do not invent a parent text. Keep a count of unresolved parent IDs in research notes so readers understand the limits of the dataset.

Exporting a second copy with all fields or all saved rows costs 0 credits. Starting a new extraction is a separate collection operation and can produce different results if comments have changed. Keep the collection date with each snapshot.