YouTube Video Details & Insights

Apify actor that scrapes public YouTube channel videos, Shorts, and direct watch links. It collects metadata (views, description, comment count, duration, published date), extracts links and keywords from descriptions, and adds computed insights. No official YouTube API key is required.

Features

• Mix up to 50 inputs in one run: channel URLs, @handles, UC channel IDs, channel names, or direct watch/Shorts URLs
• Fetch videos only, Shorts only, or both from a channel
• Up to 1000 latest items per channel
• Fast mode for bulk channel lists (channel list plus watch-page description and comment count)
• Full mode for extra metadata normalization and insights
• Direct video/Shorts URLs always use full mode and return a single item
• Built-in insights: engagement hint, comments per 1,000 views, sponsored/contact keywords, external links, content type
• Optional Apify proxy (Residential recommended) to reduce rate limits on large runs

Input

Field	Type	Default	Description
inputs	string[]	—	Bulk list (max 50 unique entries). Channel URL, @handle, UC ID, channel name, or watch/Shorts URL.
channelInput	string	—	Optional single input. Appended to inputs when both are set.
contentType	videos | shorts | both	both	What to fetch for channel inputs. Ignored for direct video/Shorts URLs.
maxVideos	integer (1–1000)	10	Latest items to process per channel. Ignored for direct links (always 1).
scrapeMode	fast | full	fast	Fast: channel list plus description and comment count. Full: extra metadata normalization and insights. Direct links always use full mode.
proxyConfiguration	object	Apify Residential	Proxy used for YouTube requests.

At least one of inputs or channelInput is required.

Example input

{
  "inputs": [
    "https://www.youtube.com/@YouTube",
    "https://www.youtube.com/watch?v=dQw4w9WgXcQ"
  ],
  "contentType": "both",
  "maxVideos": 10,
  "scrapeMode": "fast",
  "proxyConfiguration": {
    "useApifyProxy": true,
    "apifyProxyGroups": ["RESIDENTIAL"]
  }
}

Output

Results are written to the default dataset.

Video items (type: "video") include:

• Identity: videoId, title, url, channelId, channelTitle, channelHandleResolved
• Stats: views, viewsText, commentCount, commentCountText, publishedAt, duration, durationSeconds
• Text: description, shortDescription, keywords, hashtags, linksFromDescription
• Insights: engagementHint (high / medium / low / unknown), commentsPer1000Views, hasExternalLinks, hasSponsoredKeywords, hasContactKeywords, contentType (short / long / live / unknown), notableKeywords
• scrapeWarnings when a field could not be enriched

Channel summary items (type: "summary") are added after channel inputs (not after direct video URLs):

• channelId, channelTitle, channelHandle
• scrapeMode, requestedVideos, fetchedVideos, generatedAt, scrapeWarnings

Failed inputs produce an error dataset item with the message and input, so the rest of the run can continue.

Local development

Requires Node.js 20+.

npm install
npm run check
npm start

Pass input via Apify local storage (storage/key_value_stores/default/INPUT.json) or the Apify CLI. The actor reads input with Actor.getInput().

Script	Purpose
npm start	Run the actor with tsx
npm run build	Compile TypeScript to dist/
npm run check	Type-check without emitting files

Notes

• Failed channel or video inputs are skipped and recorded as error items; other inputs still run.
• Fast mode still opens each watch page for description and comment count. Full mode is similar in speed but normalizes more metadata.
• Residential proxy is recommended for large bulk runs to reduce YouTube rate limiting.