Index entire YouTube channels — video titles, descriptions, and full transcripts. Search across thousands of hours of video content in under 100ms.
✦ Transcripts auto-extracted — no YouTube API key required for public videos
Full video titles and descriptions for every video in the channel
Word-for-word transcripts extracted automatically from every video
Channel tags and video categories for better filtering
Search results link to the exact timestamp in the video
Pass any YouTube channel URL or video URL. We handle transcript extraction, chunking, and semantic indexing. No YouTube API key needed for public content.
INDEX A YOUTUBE CHANNEL
# Index MrBeast's entire channel
curl -X POST \
https://api.knowledgesdk.com/v1/projects/
{projectId}/index \
-H "x-api-key: sk_ks_..." \
-H "Content-Type: application/json" \
-d '{"url": "https://www.youtube.com/@MrBeast"}'SEARCH VIDEO TRANSCRIPTS
// Response includes timestamps
{
"results": [
{
"title": "I Spent 50 Hours Buried Alive",
"content": "At hour 32 I started...",
"timestamp": 1920,
"videoUrl": "youtube.com/watch?v=...",
"score": 0.91
}
],
"total": 24,
"ms": 67
}Index entire podcast channels so listeners can search across all episodes by topic, guest, or keyword.
Make course libraries and lecture series fully searchable. Find the exact moment a concept is explained.
Index news channels to build a searchable video archive. Find coverage of any event instantly.
Give your AI agent access to any YouTube knowledge base. Answer questions using real video content.
Challenge and philanthropy videos
Long-form interviews on AI and science
Startup advice and founder stories
Visual math explanations
Science-based health and neuroscience
Short-form dev tutorials
KnowledgeSDK automatically extracts transcripts from public YouTube videos without requiring a YouTube Data API key. Just pass any public channel or video URL.
NODE.JS SDK
import KnowledgeSDK from "@knowledgesdk/node";
const sdk = new KnowledgeSDK({ apiKey: "sk_ks_..." });
// Index Lex Fridman's channel
const job = await sdk.projects.index({
projectId: "podcast-archive" ,
url: "https://youtube.com/@lexfridman" ,
});
console.log(job.jobId);
// job_abc123 — async indexing started