Multi-agent LinkedIn content creation with GPT-4o, Google Sheets, and human review

Description: This is a Production-Grade Multi-Agent Content Engine designed for creators who prioritize technical authority over generic AI output.

Unlike standard "one-shot" prompts, this system uses a Chain-of-Thought (CoT) architecture to separate logical blueprinting from creative writing. It features a Persistent Memory loop via Google Sheets to ensure topic variety and a dedicated Compliance Editor to enforce strict branding rules (e.g., "Sentence Case" and spaced hyphens).

Built with enterprise resiliency in mind, it includes Global Error Handling and a Human-in-the-Loop Gmail approval gate with a 48-hour auto-timeout.

I built this to solve the 'Technical Content' bottleneck. It ensures that my social presence reflects the same high-quality engineering standards I apply to my client's n8n workflows.

Key features:

• Persistent Topic Memory: Automatically scans your Google Sheets history to prevent repeating technical topics.

• Multi-Agent Pipeline: Specialized agents for Topic Selection, Structural Architecture, Copywriting, and Brand Compliance.

• Style Enforcement: A final "Editor" agent that strips AI fluff and ensures a professional, human-sounding tone.

• Resilient Design: Features a global Error Trigger for instant failure notifications and an Approval Timeout to maintain system resources.

• Human-in-the-Loop: Sends a clean draft to Gmail for manual image pairing and final review before posting.

How to setup

• Google Sheets: Prepare a sheet with columns for Topic, Status, and Difficulty.

• Credentials: Connect your OpenAI (or Gemini) and Gmail accounts.

• IDs: Replace the placeholder Sheet ID and Recipient Email address in the nodes.

• Error Handling: Point the Error Trigger notification to your preferred email or Slack channel.