YouTube Video Reader

Read YouTube video metadata and authorized caption tracks

What does this node do?

The YouTube Video Reader uses YouTube Data API v3 to read metadata for up to 50 video IDs or URLs. It accepts watch, youtu.be, shorts, embed, and live URLs, preserves input order, and removes duplicates.

The node can also download a caption track when the connected YouTube account is authorized to access it. Caption download is disabled by default because it has a significantly higher quota cost than metadata reads.

Configuration

YouTube integration integration required

Google OAuth connection used for YouTube Data API requests.

videos array required

One to 50 YouTube video IDs or URLs.

Include captions boolean default: false

Explicitly enable caption discovery and download for each accessible video.

Caption language string

Optional language code, such as en or fr, used to select a caption track.

Output

The node returns its payload as text containing a JSON string — the run viewer displays proper JSON, not a Python object repr.

items array

One result per valid, deduplicated input, in input order. Each item includes success, video_id, video, transcript, transcript_status, and error.

transcript_status is available, absent, forbidden, error, or not_requested. A caption failure is isolated to its video and does not discard the other results.

Authorization and quota

Metadata uses the youtube.readonly scope. Caption download uses youtube.force-ssl and normally requires ownership of the caption track. Typical costs are 1 quota unit for videos.list or captions.list, and about 50 units for captions.download.

Warning

Enable captions only when needed. The 50-video limit and metadata-only default protect the shared YouTube API quota from unexpectedly expensive runs.

Common issues

Invalid authentication credentials or 401 error

Cause: the OAuth access token expired mid-run.

Solution: reconnect the YouTube integration in Settings > Integrations if the error persists. Otherwise, re-run the workflow — the node retries once with a refreshed token on 401, like other Google integrations.