YouTube Shorts Automation with AI-Powered Video Generation
Overview
This workflow automates the creation of engaging YouTube Shorts by leveraging AI tools to generate scripts, audio, images, and videos, then compiles them into a final video ready for upload. It uses OpenAI for script and image prompt generation, ElevenLabs for text-to-speech, Replicate for image and video generation, Cloudinary for audio storage, and Creatomate for final video rendering. The workflow segments the script into 6-second clips, creates corresponding visuals, and combines them with audio, adding smooth transitions for a polished result.
Why is this workflow useful?
Creating YouTube Shorts manually is time-consuming, requiring scriptwriting, voiceovers, and video editing. This workflow automates the entire process, enabling creators to produce high-quality, engaging Shorts in minutes. It’s ideal for content creators, marketers, or businesses looking to maintain a consistent social media presence with minimal effort.
Who is this workflow for?
- Content Creators : YouTubers or TikTok creators aiming to produce viral Shorts quickly.
- Digital Marketers : Businesses promoting products or services through short-form video content.
- Hobbyists : Individuals experimenting with AI-driven video production without advanced editing skills.
What will you need to use this workflow?
- n8n Account : Free or paid, hosted on n8n Cloud or self-hosted.
- OpenAI API Key : For generating scripts and image prompts (GPT-4o-mini model).
- ElevenLabs API Key : For text-to-speech conversion.
- Replicate API Key : For generating images and videos.
- Cloudinary Account : For storing audio files.
- Creatomate API Key : For rendering the final video.
- Basic JSON Knowledge : To understand and tweak the workflow output if needed.
How does this workflow work?
- Trigger : The workflow starts manually via the "Test Workflow" button.
- Script Generation : OpenAI (GPT-4o-mini) generates a script with an intro (40-70 chars), base (280-350 chars), and CTA (55 chars), plus a title and description with hashtags.
- Text-to-Speech : ElevenLabs converts the script into an MP3 audio file with timestamps.
- Audio Upload : The audio is uploaded to Cloudinary for storage.
- Script Segmentation : A Python script segments the audio into 6-second clips, creating transcription chunks.
- Image Prompt Generation : OpenAI generates a visual prompt for each segment in the style of CNSTLL, ensuring simple, animatable visuals.
- Image Generation : Replicate creates a 9:16 image for each segment based on the prompt.
- Video Generation : Replicate’s Minimax model generates 5-second video clips from the images.
- Video Compilation : Creatomate combines the video clips with the audio, adding fade transitions and a final black shape for a smooth close.
- Final Output : The rendered video is retrieved from Creatomate, ready for YouTube Shorts.
Setup instructions
- Import the Workflow : Copy the provided JSON into n8n and import it as a new workflow.
- Add Credentials :
- Configure OpenAI API credentials for the "Ideator" and "Image Prompter" nodes.
- Set up ElevenLabs API key in the "Script Generator" node.
- Add Replicate API key for the "Request Image," "Request Video," and "Get Video" nodes.
- Configure Cloudinary credentials for the "Upload to Cloudinary" node.
- Set Creatomate API key for the "Editor" and "Get Final Video" nodes.
- Test the Workflow :
- Click "Test Workflow" to run it manually.
- Replace
<user-query> in the "Ideator" node with your video topic (e.g., "Fitness tips for beginners").
- Verify Outputs : Check each node’s output, ensuring audio, images, and videos are generated correctly.
- Customize (Optional) :
- Adjust segment duration in the "HTTP Request" node’s Python code (default: 6 seconds).
- Modify fade duration or video resolution in the "Create Editor JSON" node.
- Save and Activate : Save the workflow and activate it for regular use.
Tips for getting the most out of this workflow
- Optimize Prompts : Use specific, concise queries in the "Ideator" node for better script quality (e.g., "Minimalist home decor ideas").
- Check API Limits : Ensure your API accounts (OpenAI, ElevenLabs, Replicate, Creatomate) have sufficient credits.
- Experiment with Styles : Tweak the CNSTLL style in the "Image Prompter" node for unique visuals, but keep prompts simple to avoid AI generation errors.
- Monitor Rendering Time : Video rendering in Creatomate may take up to 70 seconds; adjust the "Rendering" node’s wait time if needed.
- Add Hashtags : Enhance the generated description with trending hashtags like #YouTubeShorts, #AI, or niche-specific tags.
Troubleshooting
- Error in Script Generation : Verify your OpenAI API key and ensure the query is clear.
- Audio Issues : Check ElevenLabs API key and confirm the script length fits within API limits.
- Image/Video Failures : Ensure Replicate API key is valid and prompts are simple (avoid complex scenes or humans).
- Rendering Errors : Confirm Creatomate API key and check JSON structure in the "Create Editor JSON" node.
- Contact Support : Reach out to n8n Community Forum or respective API providers for persistent issues.
Built with
- n8n Nodes : Manual Trigger, OpenAI, HTTP Request, Set, Split Out, Aggregate, Merge, Wait.
- External Services : OpenAI (GPT-4o-mini), ElevenLabs (Multilingual v2), Replicate (Flux-Cinestill, Minimax Video-01), Cloudinary, Creatomate.
- Custom Code : Python for audio segmentation and JSON rendering.
Example output
For a query like "Morning routine hacks":
- Script :
- Intro : "Struggling to start your day right?"
- Base : "Try these 3 morning hacks: 5-min meditation to clear your mind, a quick stretch routine to boost energy, and a cold shower to stay alert. Set your day up for success!"
- CTA : "Save this and try it tomorrow!"
- Title : "3 Morning Hacks You NEED to Try!"
- Description : "Kickstart your day with these easy hacks! 🧘♀️🚿 #MorningRoutine #Productivity #YouTubeShorts"
- Video : A 15-second Short with three 5-second clips (meditation scene, stretching POV, cold shower close-up), synced with audio and fade transitions.
