Showcase: Omni-Channel AI Publisher & Strict Legal Compliance Engine (n8n Cloud + Claude 3.5 Sonnet)

Built with n8n
1

Disney workflow
Disney workflow1920×793 148 KB

Hey community,

I wanted to share a production-grade content engine I recently architected to solve a major issue for scaling digital agencies: Automating multi-channel social media publishing while maintaining strict legal brand compliance.

When agencies scale content across Facebook, Instagram, and TikTok, human oversight often creates an operational bottleneck—especially when dealing with strict vendor compliance guidelines (like protected trademarks, restricted nomenclature, and mandatory disclaimers).

Here is a high-level breakdown of how this unified system handles it autonomously right on n8n Cloud:

:gear: The Architecture Components:

  • Time-Staged Polling: An automated Schedule Trigger pulls raw promotional parameters from a master queue, isolating rows tagged as pending.
  • Algorithmic Asset Selection: A custom JavaScript Code Node executes a deterministic math loop (Math.floor(Math.random())) to dynamically select random visual asset urls from a storage array, preventing programmatic content fatigue.
  • High-Constraint Context Engineering: Instead of a basic completion, this pipeline binds strict brand rules directly into an Advanced AI Agent Node paired with Claude 3.5 Sonnet. The agent evaluates input data against strict legal gates, preventing trademark misuse before it happens.
  • Human-in-the-Loop Slack Stop: Uses the asynchronous sendAndWait execution node. The system pauses mid-run, generates an approval link, and pushes it to Slack. It holds the runtime memory until a human stakeholder clicks a button, ensuring zero unvetted posts go live.
  • Parallel Execution & State Reset: Fires simultaneous, type-safe API payloads to the platforms before executing a trailing Update Row Google Sheets connector to update database states and prevent duplicate runs.

:hammer_and_wrench: How to Inspect the Build:
I don’t use traditional code repositories like GitHub, so I have documented this entire layout, along with the complete clean JSON workflow tree and the custom JavaScript scripts, in a public Notion workspace:

:backhand_index_pointing_right: :rocket: Mac McCoy’s AI Portfolio

Feel free to copy the JSON block inside my portfolio and paste it directly onto your own canvas to inspect the node mapping, model setups, and data handling structures!

I am a specialized AI Automation Contractor currently open to new fractional or project-based B2B client builds within European timezones. If your agency is facing an optimization bottleneck, feel free to drop me a message here or connect via the links in my portfolio.