wtf is gp-wf-005-youtube-transcript-cleaner?

This is a small automation workflow for n8n that takes a rough, messy transcript, such as one pulled from a YouTube video, and turns it into a clean, readable piece of Markdown text while keeping the original meaning intact. You feed it details like the source title, what kind of content it is, the raw transcript, your goal for the content, your audience, and how thorough the cleanup should be. The workflow is part of a free library of starter n8n workflows meant to be simple enough to read through, useful enough to actually run, and easy enough to adjust. The author points out that raw transcripts are hard to read, quote, search, or reuse, and this tool exists to fix that without becoming a giant, complicated automation system. Behind the scenes, your inputs are sent to a chat model running locally through Ollama. The AI's response is then reformatted by some JavaScript steps inside the workflow, and the final result is written into a Google Doc containing the cleaned transcript itself, key takeaways, ideas for repurposing the content elsewhere, notes on what was changed during cleanup, and flags for anything that might need a human review. To use it you need n8n installed, Ollama running locally with a chat model such as llama3.1:8b, and a connected Google Docs credential inside n8n. Setup means importing the workflow file, hooking up your Google account, and making sure the model name matches what is actually installed locally. No private credentials are bundled with the workflow, and it is released under the MIT license.