Publishing a YouTube video is only the beginning. The real leverage comes from repurposing that content across multiple different platforms.

One effective way is to convert your video transcript into a polished LinkedIn post and newsletter.

And with the right tools, N8N and Apify, you can automate the entire process.

This article walks through a practical workflow which detects new YouTube video uploads and extracts the transcript. Then, it generates a newsletter draft using Claude and saves it to Google Docs for review and publishing.

How the Workflow Operates

The automation consists of six connected nodes inside n8n, each handling a distinct step in the pipeline. Once configured, the workflow runs on a schedule and requires no manual intervention until the final review stage.

Step 1: Detect New YouTube Uploads with RSS

The workflow begins with an RSS feed trigger. Instead of polling the YouTube API directly, this approach uses the YouTube channel RSS feed. This method provides a lightweight and reliable way to detect new uploads.

To set this up, first add an RSS Feed Trigger node in n8n. Then, pass in your YouTube channel’s feed XML URL and include your channel ID. The trigger can be configured to check every 10, 30, or 60 minutes depending on your upload frequency. Each time it fires, it returns the latest video title, link, and metadata.

Step 2: Extract the Transcript with Apify

The workflow identifies a new video. It then passes the video URL to an Apify node. This node is configured with the YouTube Transcript Scraper actor.

This actor pulls the full transcript of the video, including timestamps. You will need an Apify account and a saved actor configuration. In n8n, connect the Apify node and pass in the video URL from the RSS feed output. The node waits for the scraping job to complete before moving forward.

A second Apify node then retrieves the dataset items, which contain the raw transcript segments.

Step 3: Clean and Prepare the Transcript

The raw transcript arrives as multiple items, each containing a text segment and a timestamp. Since the timestamps add noise and are not useful for a newsletter, the next steps strip them out.

A Split Out node separates each transcript segment into individual items. An Aggregate node then recombines them, pulling only the text fields and merging them into a single clean list. The result is one continuous block of transcript text, ready to be passed to a language model.

Make sure to enable the merge list option in the Aggregate node. Without it, the text remains fragmented.

Generating the Newsletter with Claude

The cleaned transcript feeds into a Claude node configured to message a model. The prompt instructs Claude to act as a professional writer creating a LinkedIn newsletter for a broad business audience.

Key elements of the prompt include:

– Writing a complete 400-word newsletter

– Explaining ideas clearly at roughly a sixth-grade reading level

– Maintaining a professional and accessible tone

The full transcript is appended at the bottom of the prompt. Enable simplify output on the Claude node to return only the generated text without extra metadata.

You will need a Claude API key, which you can generate from Anthropic’s API console. Enter the key in n8n’s credential manager to authenticate the node.

Saving the Output to Google Docs

The final stage creates a Google Docs document and writes the newsletter content into it. This requires two nodes: one to create the document and one to update it with the generated text.

Before this works, you need to configure credentials in Google Cloud Console. Enable both the Google Docs API and the Google Drive API for your project. After that, create OAuth credentials. Then add them to n8n. Set up the OAuth consent screen as well.

The workflow saves the document to a specified Google Drive folder. From there, you can review, edit, and manually publish the newsletter to LinkedIn.

Watch the full step-by-step tutorial below and join my Skool community here.