YouTube Channel RSS Scraper - Recent Video Feeds

Read recent videos from YouTube’s public channel RSS feeds. Export titles, descriptions, publication dates, links and available feed metrics, with keyword, age and numeric filters. This is a recent-feed monitor, not a complete channel-history scraper.

Try a small export: Linus Tech Tips recent-video feed example. Free-plan result events cost $20 per 1,000 video records, plus the Actor-start event. Check the Pricing tab for your account’s discounts.

Tested input

{
  "channelIds": [
    "UCXuqSBlHAE6Xw-yeJA0Tunw"
  ],
  "maxResults": 3
}

Run the Actor, open Dataset → Overview, then export JSON, CSV or Excel. The example above returned 3 records in a September 10, 2026 test run. Source availability and ordering can change.

Real output

Selected fields from that run:

{
  "video_id": "sL6OWsT47zc",
  "title": "Apple just Copied Microsoft’s Worst Idea - iPhone Duo, Airpods 5, iPhone 18 Pro & More",
  "channel_title": "Linus Tech Tips",
  "view_count": 754604,
  "rating_count": 21253,
  "published_at": "2026-09-10T08:28:52+00:00",
  "viral_score": 94,
  "url": "https://www.youtube.com/watch?v=sL6OWsT47zc"
}

Input controls

Supply 1–100 channelIds beginning with UC; URLs and handles are not accepted here. Duplicate IDs run once. maxResults is a global export limit from 1 to 1,000. keywords matches any supplied term in the title or description. minViews, minRatings, minAverageRating and maxDaysOld filter the recent feed; they do not make YouTube return older entries. Set includeViralScore: false to omit the score.

Scope and field availability

RSS feeds expose only a recent window, commonly about 15 videos per channel. Raising maxResults cannot recover full history. The source channel_id may omit the UC prefix; use channel_url to retain the canonical channel reference. Missing source metrics remain null. Feed rating fields are not a complete engagement report.

Costs, limits and failures

On the Free plan, each exported video record costs $0.02. The start event costs $0.00005 per allocated GB, with a minimum of one event. At 512 MB, the 3-record example costs $0.06005 in Actor events before discounts. Prices remain unchanged in this update.

Set a maximum charge in the run options. Delivery stops when the result budget is reached. Duplicate input URLs or IDs are collapsed before retrieval; rerunning the same input later is a new billable observation. Limits are upper bounds, not guaranteed row counts.

The Output tab links to SUMMARY. Source errors fail the run while retaining any valid rows already exported; successful earlier rows retain their result charges. Inspect the summary, dataset and run log before retrying. An unavailable source is not evidence that a channel, playlist or video was deleted. A legitimate empty collection or a filter excluding every video can still return zero rows successfully.

Export an existing run with Python

Install apify-client, set APIFY_TOKEN and APIFY_DATASET_ID, and export the completed dataset without starting another run:

import json
import os
from apify_client import ApifyClient

client = ApifyClient(os.environ["APIFY_TOKEN"])
rows = list(client.dataset(os.environ["APIFY_DATASET_ID"]).iterate_items())
with open("youtube-channel-intelligence.json", "w", encoding="utf-8") as output:
    json.dump(rows, output, ensure_ascii=False, indent=2)

Save a Task for each recurring input and schedule it through Apify. API and automation workflows can reuse the same input. Compare video IDs between runs when building a monitoring database.

Questions before running

Does the score predict virality? No. The existing viral_score field is a deterministic 0–100 calculation using current feed views, ratings and video age. Missing metrics contribute as zero to the calculation. It is neither a trained prediction nor evidence of future performance.

Are results sorted by score? No. Results retain feed and input-channel order; sort the exported rows yourself if needed.

Why are there fewer than 1,000 videos? The cap applies after filtering a short RSS window. This Actor does not paginate a channel’s entire catalog.

Related YouTube workflows

• Keyword search discovers video URLs from phrases.
• Video metadata enriches known individual video URLs.
• Transcripts and subtitles exports available caption tracks.