Build a workflow that makes a video from a script and uploads it unlisted
Idea + starter prompt, not run by us yet
For: Creators who want to see how a script-to-upload video pipeline fits together
An n8n workflow starts from a Telegram message or a schedule. An AI model writes a JSON script, ElevenLabs voices each segment, Pexels supplies footage and Shotstack renders it. The video goes up to YouTube as unlisted so you can check it first, and Telegram and the sheet are updated. A 10-minute video took about 3 to 5 minutes to render.
Steps
- Sketch the stages on paper first: script, voiceover, footage, render, upload, log. Build and test one stage at a time.
- Ask Claude Code to help you wire each n8n node and explain what it does. Use free or trial keys kept in n8n credentials.
- Keep uploads as unlisted so you review every video before it goes public.
- Plan for YouTube's API quota. The author notes it allows only around six uploads a day, so test with short clips first.
- Check the finished video yourself for wrong facts and bad footage before you publish.
Starter prompt
Trial first: plan one 30-second test video on a harmless topic, [test topic], and set any upload to private on a test channel, never my main one. Log the trial in a separate sheet tab called [trial]. Help me plan an n8n flow. It starts when a topic arrives by [Telegram or a schedule], writes a short script as JSON, makes a voiceover, picks stock footage, renders the video, uploads it to YouTube as unlisted and logs the result in a Google Sheet. Draw it as a list of nodes with inputs and outputs. Name the risks, including API quotas and wrong facts in the script. Stop and wait for my yes before each upload and before any step that spends credits or API quota. Showing me the render is not permission. My API keys stay in n8n credentials, so do not ask me to paste them. Never make anything public.
Original post: community.n8n.io · found via agentusecases.ai. Written up in our own words. We haven't run it, so test it on made-up data first.