Repurpose videos into posts
One video holds a blog post, a newsletter, a thread and a week of short posts. The bottleneck is getting the words out of the video. With a transcript API that step takes a second, so the whole pipeline can run on its own every time you publish.
The workflow
- Trigger: a new upload on your YouTube channel (
/v1/videoson a schedule), a new podcast episode in your feed, or a link pasted in Slack. - Transcript:
/v1/transcriptwithformat=text. Works for YouTube, TikTok, Reels and podcasts. - Write: one LLM call per format, each with its own prompt and examples of your voice.
- Review and schedule: drafts go to Notion, Google Docs or your scheduler for a human pass.
Prompts that work
| Format | Prompt idea |
|---|---|
| Blog post | "Rewrite this transcript as a 1,200-word article with H2 sections. Keep the speaker's examples and numbers. Remove filler and repetition." |
| X / LinkedIn thread | "Extract the 5 most surprising claims and write each as a standalone post under 280 characters, with the timestamp." |
| Newsletter | "Summarize in 3 paragraphs for people who didn't watch, then list 3 takeaways and a link to the video." |
| Short clips | "From the timestamped transcript, pick 3 segments of 30 to 60 seconds that work alone, with start and end times." |
| Subtitles | Skip the LLM: request format=srt and upload the file. |
Why timestamps matter
Ask for the JSON format and every segment comes with its start time. The LLM can then point to the exact moment for a clip, a quote card or a "watch from here" link, and you can check any claim against the source in a click.
Build it in n8n
Schedule Trigger, Transkrio List Videos, Transkrio Get Transcript, then an OpenAI or Anthropic node per format and a Notion node for drafts. Details on the n8n page. Prefer chatting? Connect the MCP server to Claude and ask "turn this video into a LinkedIn post". To try one video by hand, use the free transcript tool.