Go to example tasks
Build a RAG knowledge base from ML lecture videos
Created by
Omar Nagy
Turns three foundational ML lectures (3Blue1Brown on neural networks and on attention, Karpathy building GPT from scratch) into embedding-ready chunks of about 400 tokens with 40 tokens of overlap. Every chunk carries its start timestamp, video id and title, so an answer can cite the exact moment it came from. About $0.02 a run. Needs residential proxy access on the plan that runs it.
YouTube Transcripts for RAGomargnagy/youtube-transcripts-for-rag
Chunk text
Chunk ID
Chunk index
Tokens
+10 fieldsTextNumberBooleanListObject
Input
YouTube videos(required):https://www.youtube.com/watch?v=aircAruvnKk+2
Mode:chunks
Chunk size in tokens:400
Chunk overlap in tokens:40
Include the full transcript on the video record:true
Language fallback chain:en
Proxy
Residential bandwidth budget (MB per run):200
Output fields
Chunk text
Chunk ID
Chunk index
Tokens
Start
End
Deep link
Title
Channel
Video ID
Language
Auto-generated
Segments
Record type
Sign up on Apify01
Create your Apify account to access the YouTube Transcripts for RAG.
Start the run02
The Actor will start running based on the input automatically.
Receive the output03
Monitor the progress in real-time. You will be notified as soon as your dataset is complete and ready for review.
Integrate into your workflow04
The final output is delivered in JSON, CSV, or Excel format, ready to be plugged into your workflow.
