YouTube New Comment Monitor and Filter

Use this YouTube comment monitor to emit newly observed top-level comments from known public videos. Schedule repeat runs to create a bounded new-comment feed, then narrow results with exact phrases or an ISO timestamp.

This is an unofficial public-page tool and is not affiliated with YouTube.

Quick start

1. Add one or more public YouTube video URLs.
2. Choose a maximum number of returned comments per video.
3. Run once to save observed comment IDs.
4. Run again to receive comments not observed by the earlier runs.

{
  "video_urls": ["https://www.youtube.com/watch?v=jNQXAC9IVRw"],
  "max_comments_per_video": 100,
  "exact_phrases": ["pricing"],
  "published_after": "2026-01-01T00:00:00Z"
}

YouTube comment filters and exports

• exact_phrases requires every supplied case-sensitive phrase to appear.
• published_after accepts an ISO-8601 timestamp and filters comments when YouTube exposes a parseable publication time.
• Results contain comment ID, text, author text, source URL, and publication time when available. Export them through Apify datasets or the API.

How repeat runs work

The Actor stores observed comment IDs in an Actor-owned state store. A comment is emitted at most once for the same monitored video while that state remains.

Limitations

• Reads public comments only and does not accept login cookies or credentials.
• Collects one public continuation response per video rather than every comment.
• Returns top-level comments; reply threads are not collected.
• YouTube can change or withhold its public continuation data at any time.
• Relative publication labels may not support ISO-date filtering.

Responsible use

Comment text and author labels can be personal data. Follow applicable law, YouTube's terms, and reasonable rate limits when storing or using results.

Trial and support

Start with one video and a 20-comment limit, leaving phrase and date filters empty. Expect top-level comment records when new observations are available. A later run can return only a run_summary with zero items. State is shared by video across this Actor's runs in your account; no new records is not proof of a broken run. Check logs for source or state diagnostics. Check the Pricing tab and run cost estimate before starting. Charges may apply even when no new comments are returned. Use the Issues tab for questions.

Result billing

The new pricing contract is $0.0005 per start at 256 MB plus $0.15 per 1,000 delivered new comments, on every Apify plan. Check the Pricing tab for the currently effective contract and any scheduled change. Platform usage is included in event prices.

Diagnostics and unchanged job/comment summaries have no result charge; the startup fee still applies. A successful website check is billable even when the page is unchanged. If the run budget cannot cover the result batch, no result rows are delivered or billed and the baseline is not advanced. Increase the run limit and retry. RUN_REPORT in the run's key-value store records diagnostics and delivery status. Failed or interrupted deliveries may be repeated on a later run; use record IDs to deduplicate.

If every video source fails, the run fails and retains its previous baseline. RUN_REPORT identifies the failed request stage and extraction status. A successful check with no new comments remains a successful run.