youtube

YouTube

Access YouTube channel data, videos, search, and transcripts.

Requires Rebyte API auth — $AUTH_TOKEN and $API_URL are set up per the agent's system prompt; use them as Bearer token and base URL.

Data Proxy Operations

All official YouTube API operations go through the data proxy. Use $API_URL and $AUTH_TOKEN from the authentication setup above.

Get Channel Info

Look up a YouTube channel by handle to get its metadata and uploads playlist ID.

curl -X POST "$API_URL/api/data/youtube/channel" \
  -H "Authorization: Bearer $AUTH_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"handle": "mkbhd"}'

Response:

{
  "success": true,
  "channelId": "UCBcRF18a7Qf58cCRy5xuWwQ",
  "name": "Marques Brownlee",
  "uploadsPlaylistId": "UUBcRF18a7Qf58cCRy5xuWwQ",
  "statistics": {
    "subscriberCount": "19600000",
    "videoCount": "1800"
  }
}

Use uploadsPlaylistId with the playlist items operation to get the channel's videos.

Get Playlist Items (Latest Videos)

Fetch videos from a playlist. Use with uploadsPlaylistId to get a channel's latest uploads.

curl -X POST "$API_URL/api/data/youtube/playlistItems" \
  -H "Authorization: Bearer $AUTH_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "playlistId": "UUBcRF18a7Qf58cCRy5xuWwQ",
    "maxResults": 10
  }'

Parameters:

• playlistId (required): Playlist ID (use uploadsPlaylistId from channel response)
• maxResults (optional): 1-50, default 15
• pageToken (optional): For pagination

Response includes videos[] with: videoId, title, description, publishedAt, url

Get Video Details

Get duration, view count, like count for specific videos. Useful for filtering out Shorts.

curl -X POST "$API_URL/api/data/youtube/videoDetails" \
  -H "Authorization: Bearer $AUTH_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"videoIds": "dQw4w9WgXcQ,jNQXAC9IVRw"}'

Parameters:

• videoIds (required): Comma-separated video IDs (max 50)

Response includes videos[] with: duration, isLikelyShort, statistics, caption (whether captions exist)

Search YouTube

Search for videos, channels, or playlists.

curl -X POST "$API_URL/api/data/youtube/search" \
  -H "Authorization: Bearer $AUTH_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "q": "machine learning tutorial",
    "type": "video",
    "maxResults": 10,
    "order": "relevance"
  }'

Parameters:

• q (required): Search query
• type (optional): video, channel, or playlist (default: video)
• maxResults (optional): 1-50, default 10
• order (optional): relevance, date, rating, viewCount, title
• channelId (optional): Filter to a specific channel
• pageToken (optional): For pagination

Note: For a channel's latest videos in chronological order, use channel + playlistItems instead of search.

---

Transcripts

The official YouTube API does not support downloading transcripts for videos you don't own. Use the youtube-transcript-api Python library instead.

Setup

pip install youtube-transcript-api

Get Transcript

from youtube_transcript_api import YouTubeTranscriptApi

ytt_api = YouTubeTranscriptApi()
transcript = ytt_api.fetch("VIDEO_ID")

# Combine into full text
full_text = " ".join(segment.text for segment in transcript)
print(full_text)

Get Transcript with Timestamps

ytt_api = YouTubeTranscriptApi()
transcript = ytt_api.fetch("VIDEO_ID")

for segment in transcript:
    print(f"[{segment.start:.1f}s] {segment.text}")

Notes

• Add 2-second delays between requests to avoid rate limiting
• Some videos don't have transcripts (live streams, music, disabled captions)
• Auto-generated transcripts may have spelling errors for names and technical terms
• Use videoDetails operation first to check if caption is true before attempting transcript fetch

---

Common Workflows

Get Latest Videos from a Channel

1. Call channel with the handle to get uploadsPlaylistId
2. Call playlistItems with that playlist ID
3. Call videoDetails to filter out Shorts (isLikelyShort: true)

Get Transcript of a Video

1. Call videoDetails to confirm captions exist
2. Use youtube-transcript-api to fetch the transcript

---

Delivering Output

After generating any output files (transcripts, reports, etc.), upload them to the Artifact Store so the user can access them.