Skip to content

Latest commit

 

History

2 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 

Repository files navigation

⚙️ HelioConvert Autonomous Content Engine

An intelligent, fully automated n8n workflow designed to transform raw technical concepts into highly opinionated, zero-fluff LinkedIn posts. It features algorithmic angle rotation, strict AI copywriting constraints, and dynamically generated branded graphics—allowing for continuous publishing without human intervention.

🧠 Architecture Overview

This "Content Generator Brain" runs on a schedule to fetch raw concepts from an Airtable Knowledge Vault. It uses Google Gemini to draft technical, anti-buzzword content, evaluates the text to determine the optimal visual format, and generates a custom HTML/CSS graphic on the fly. The final payload is automatically published to LinkedIn and logged back into Airtable.

✨ Key Features

  • Knowledge Vault Extraction: Pulls topics, pillars, core contexts, and key takeaways directly from a structured Airtable database.
  • Dynamic Angle Rotation: Randomly assigns one of five distinct writing angles (Architectural Breakdown, Bottleneck & Cost of Inaction, Contrarian / Unpopular Take, Client Transformation Story, or Quick Strategic Tip) to keep the feed unpredictable.
  • Opinionated AI Copywriting: Utilizes Google Gemini with strict negative prompting (banning AI buzzwords like "delve" or "landscape," forbidding generic hooks, and enforcing a blunt, technical tone).
  • Algorithmic Visual Routing: A secondary AI router analyzes the draft and classifies it into a visual archetype (flowchart, metric_dashboard, split_compare, tech_stack, or text_only).
  • On-the-Fly Image Generation: Injects extracted copy into dynamic HTML/CSS templates (featuring Slate, Glows, and Glassmorphism styling) and converts them to high-quality images via the HTML2Img API.
  • Automated Distribution & Logging: Publishes text or image posts directly to LinkedIn and updates the Airtable Content Queue with the final AI draft and "Published" status.

🛠️ Tech Stack

  • Orchestration: n8n
  • Database: Airtable
  • LLM / AI: Google Gemini (via Langchain nodes)
  • Image Generation: HTML/CSS/JS + HTML2Img API
  • Distribution: LinkedIn API

🚀 Workflow Breakdown

  1. Schedule Trigger: Initiates the workflow three times daily (05:47, 08:22, 12:13).
  2. Data Fetching: Searches the Helio Convert Publishing Engine Airtable base.
  3. Angle Selection (JS Code): Randomizes the content approach.
  4. Draft Generation (Gemini): Authors the post based on strict technical constraints.
  5. Visual Routing (Gemini): Parses the generated draft and outputs a JSON config dictating the visual structure and UI copywriting.
  6. Conditional Formatting (If Node):
    • If Image: Generates a custom HTML template, renders it into a 1080x1080 image, downloads the file, posts to LinkedIn, and creates a published record in Airtable.
    • If Text-Only: Bypasses image generation, posts the raw text to LinkedIn, and updates the Airtable queue.

⚙️ Installation & Setup

  1. Import the provided JSON workflow file into your n8n instance.
  2. Configure your credentials for:
    • Airtable API
    • Google Gemini (PaLM) API
    • LinkedIn OAuth2 API
  3. Update the HTTP Request node with your active HTML2Img API key.
  4. Ensure your Airtable base matches the table structure (Topic/system name, Pillar, Core Context & Metrics, Key Takeaway/Lesson).
  5. Activate the workflow!

Author: Ginger-Eke Chienyegom
B2B Digital Systems Architect & Founder, HelioConvert