AI Content Pipeline (LinkedIn + X + Email)
One scheduled trigger generates a LinkedIn post, X post, image prompts, and a research-backed email — all saved to Supabase automatically every 3 days.
The Problem
Posting quality social media content consistently is exhausting and time-consuming. Most people open a blank doc, stare at it, rewrite the hook, second-guess the structure, and spend hours on a single post. When that has to happen every 3 days with content that is credible, research-backed, and still feels human, the process completely breaks down and never scales.
What I Built
Built a fully automated content pipeline using n8n, GPT-4o, and a RAG knowledge base. One scheduled trigger pulls from domain-specific material so the output actually sounds like the person it represents — not generic AI fluff. It writes a LinkedIn post in the right tone, generates an X post under 280 characters, creates image prompts ready to use, tracks research sources internally, saves all content to Supabase, delivers a full email with everything inside, and closes the schedule row automatically. The topic is defined once in a spreadsheet. The system handles everything else.
The Result
Content produced every 3 days with zero manual writing. Output is domain-specific and human-sounding, not surface-level AI takes. The human bottleneck is fully removed — no brief writing, no drafting, no schedule managing.