How to Repurpose One Long YouTube Video Into 10 Social Media Posts Using NotebookLM

Repurposing a long YouTube video into bite-sized social content usually requires jumping between transcription tools, ChatGPT, and document editors.

With NotebookLM’s massive 1-million-token context window and its powerful Gemini 3.5 engine, you can drop a video link or transcript into a single notebook and instantly generate an entire ecosystem of social assets.

Here is the exact step-by-step workflow to turn one long video into 10 platform-optimized social media posts (4 Tweets/X posts, 4 LinkedIn posts, and 2 Emails).

1 Create Your Source Material

2 mins

1.Create Your Source Material:2 mins.

Open NotebookLM and create a new notebook. Click Add Source and select YouTube. Paste your video link.

Note: If the automated YouTube transcript fails to load or is unavailable, use a free tool like DownSub to grab the .txt file, and upload it as a text source instead.

2 Configure Your Custom Persona

1 min

2.Configure Your Custom Persona:1 min.

Open NotebookLM’s Chat Settings and enable a Custom Chat Persona. Instruct the AI to act as a Head of Digital Growth and Copywriting Expert. This forces the engine to bypass generic, overly technical text and generate engaging, high-conversion copy.

3 Run the Core Asset Prompts

5 mins

3.Run the Core Asset Prompts:5 mins.

Instead of asking for all 10 posts at once (which dilutes the quality), use specific, targeted prompts in the chat panel to batch out your platform assets sequentially.

4 Export and Schedule

2 mins

4.Export and Schedule:2 mins.

Review the outputs in the chat panel. Click the Save as Note icon to instantly turn them into structured workspace notes, or export them directly as a Markdown/Docx file to drop into your scheduling tool.

To get the absolute best results, run these three prompts one by one in your notebook chat. NotebookLM will scan the entire transcript to extract distinct, non-overlapping hooks and takeaways.

1. The X (Twitter) Batch: 4 Specific Styles

Prompt: Based on the uploaded video transcript, extract 4 distinct key insights and write 4 highly engaging X (Twitter) posts. Vary the formats exactly as follows: Post 1: A contrarian hook or counter-intuitive take. Post 2: A clear “How-to” list breaking down a process. Post 3: A “Before vs. After” transformation story based on the data. Post 4: A punchy one-line value statement followed by a 3-bullet breakdown. Keep all posts under 280 characters, use clean spacing, and avoid generic hashtags.

2. The LinkedIn Batch: 4 Authority-Building Posts

Prompt: Analyze the video source and generate 4 long-form LinkedIn posts tailored for a professional audience. Structure them to prioritize high-hook readability: Post 1: A mistake-driven narrative (“The #1 mistake people make when…”). Post 2: A comprehensive framework breakdown (Step-by-step tactical value). Post 3: A perspective shift challenge (Challenging standard industry norms). Post 4: A contrarian insight backed by data or direct examples from the video. Ensure each post starts with a strong 1-line hook, uses short paragraphs, and ends with a conversation-starting question.

3. The Email Newsletter Batch: 2 High-Value Broadcasts

Prompt: Using the transcript data, write 2 value-first email newsletters for an audience of subscribers. Email 1: A “Curiosity Gap” email—open with a compelling problem mentioned in the video, explain the lesson learned, and tease the underlying strategy. Email 2: A “The Ultimate Guide” email—provide a highly scannable, deeply tactical breakdown of the core concept. Include clear Subject Line options and a natural placeholder for a Call to Action (CTA) link.

  • Strict Fact Grounding: Unlike traditional AI models that hallucinate generic advice when writing social copy, NotebookLM relies strictly on the source material. Every post it generates will use your exact frameworks, stories, and terminology. Google NotebookLM
  • Built-in Citations: If a generated post references a specific statistic or case study from your video, NotebookLM maps a dynamic citation number next to it. Clicking it brings you to the exact section of the transcript to verify its accuracy.
  • Zero Context Loss: With its deep reasoning architecture, it safely retains the nuance of a 3-hour podcast or a 10-minute tutorial without losing thread clarity half-way through your prompt sequence.