Turn long YouTube playlists into fast answers and clear summaries. Content and marketing teams can research videos without watching them all. Paste a playlist or video link and get insights, topics, and answers in minutes.

A chat entry point collects your link and routes by intent to either playlist or single video processing. The flow pulls video lists and transcripts, merges titles and text, and creates structured summaries. It splits the text into chunks, builds embeddings with Google Gemini, and stores them in Qdrant for search. Redis keeps session context, so follow up questions use the same data. A retrieval step reads from the vector store to answer new questions about the playlist or video. An AI agent also writes a full summary per video for quick reading.

You will need access to Google Gemini, YouTube Data API, Qdrant, and Redis. Expect to cut hours of viewing and note taking into a short chat session, and control scope with a simple limit on how many videos to process. Great for campaign planning, competitive research, topic analysis, and training content review.

• Chat trigger prompts users to paste a YouTube playlist or video link and starts processing
• Intent routing detects if the link is a playlist or a single video and sends it down the right path
• Playlist limit asks for how many videos to process and applies a hard cap
• HTTP requests to YouTube Data API fetch video lists, titles, and details for each item
• Transcript fetch and concatenation groups text by video and prepares it for analysis
• Structured summary prompts create clear topic outlines and key insights per video
• Embeddings with Google Gemini and chunking store content in Qdrant for fast retrieval
• Redis memory holds session context and a delete step clears collections for a clean reindex

• Reduce manual viewing from 5 hours to 15 minutes for a typical playlist
• Automate up to 80 percent of transcript review and note taking
• Improve research accuracy by pulling answers from verified transcript text
• Handle 10 times more videos with chunking and vector search
• Connect Google Gemini, YouTube, Qdrant, and Redis in one flow
• Keep chat context across questions so you do not repeat work

1. Import the template into n8n: Create a new workflow in n8n > Click the three dots menu > Select 'Import from File' > Choose the downloaded JSON file.
2. You'll need accounts with Google Gemini, Qdrant, Redis and YouTube Data API. See the Tools Required section above for links to create accounts with these services.
3. In the n8n credentials manager, create a Google Gemini credential. If prompted, click Create new credential and follow the on screen steps. Use your API key from Google AI Studio. In Gemini nodes, select models like models/gemini-2.0-pro-exp for chat and models/text-embedding-004 for embeddings.
4. Create a Qdrant credential. Use your Qdrant Cloud URL and API key. In Qdrant nodes, pick this credential and keep the collection name dynamic as configured so it matches the current playlist or video id.
5. Add a Redis credential. Enter host, port, and password from your Redis provider. Test the connection. The context nodes will use this to store and read session data.
6. Enable the YouTube Data API v3 in Google Cloud and create an API key. Open the Playlist HTTP Request and Video HTTP Request nodes and add the key in the query string or header. Verify endpoints for playlistItems and videos are set correctly.
7. Open the Chat trigger node, set it to public if needed, and copy the webhook URL. Activate the workflow so the chat UI can accept links.
8. Review the limit setting in the Numb of Videos and Playlist Limit nodes. Set a default that fits your quota and use case.
9. Check the embedding and chunking settings in the text splitter and embedding nodes. Default chunk size 1200 and overlap 200 are set for balanced context and cost.
10. Run a test: paste a playlist URL in the chat, enter the number of videos, and watch the status nodes. Confirm a new Qdrant collection is created and that answers come from the vector store nodes.
11. Troubleshoot common issues: if transcripts are missing, the video may be private or transcripts disabled. If you hit API rate limits, lower the video limit. If stale data appears, trigger the Delete Collection step and reindex.