I recently built an AI TikTok / Shorts Content Factory prototype using n8n.
The idea behind this workflow is simple:
A user submits one content idea, and the workflow turns it into a structured short-form content production pipeline with AI generation, human approval, tracking, and ready-to-post status.
What the prototype includes:
โ
Content input form
โ
AI-generated hook, script, captions, scene plan, and prompts
โ
Airtable content tracking
โ
Human approval before moving forward
โ
Approve / reject / revise routing
โ
Revision Agent for improving content based on feedback
โ
Second approval after revision
โ
Final video prompt preparation
โ
Voiceover placeholder for future voice API integration
โ
Video generation placeholder for production API integration
โ
Final video review
โ
Posting placeholder
โ
Final Airtable status: Ready to Post
This is currently a working prototype, not a full production system yet.
The real voiceover, video generation, and posting APIs can be connected depending on the clientโs tools, API access, and production requirements.
What I focused on most was building the workflow architecture properly:
Idea โ AI content โ Airtable tracking โ human approval โ revision โ final prompt โ voice/video/posting stages.
I believe human approval is important in AI automation because it keeps quality control in the process instead of blindly automating everything.
Special offer for first clients:
Iโm opening a few discounted MVP build spots for early clients who want to test AI automation in their content or business workflow.
This can include:
โข AI content workflows
โข Email reply agents
โข CRM / lead follow-up automations
โข Human approval workflows
โข Airtable + n8n automation systems
The goal is to build a practical working prototype first, then improve it based on the real use case.
Open to feedback from anyone building AI content workflows, short-form video systems, or client automation workflows.