LinkedIn Content Engine: Automated creation & publishing
Full n8n system for LinkedIn content generation and publishing: real-time web research, multi-framework AI writing, DALL·E image generation, and scheduled posting, all driven from Airtable with human approval.
Cadence
Writing frameworks
Human validation
Context
Founders and consultants need to maintain a consistent LinkedIn presence, but writing 3 posts a week takes hours (research, writing, visuals, scheduling). Many give up or delegate to expensive agencies that fail to capture their voice.
The challenge
Build an agent capable of generating 3 LinkedIn posts per week while respecting the brand voice, varying writing frameworks, doing real-time web research to back up points, generating theme-coherent images, and publishing automatically after human approval.
The solution
An n8n workflow in 2 stages: (1) a daily 7am scheduler that pulls topics from Airtable, does web research with gpt-4o-mini-search-preview, randomly picks 1 of 4 frameworks (PAS / Insight-Impact-Recommendation / Story-Insight-Action / What-So What-Now What), writes with gpt-4o + brand guidelines, generates a DALL·E image if needed, then sets the post to 'Needs Approval'. (2) A second scheduler Mon/Wed/Thu at 7:45am that publishes 'Approved' posts via LinkedIn API.
Key features
- 01Daily 7am schedule trigger for automatic content generation
- 02Airtable filtering of 'Waiting for Content' topics tagged LinkedIn
- 03AI Query Generator: 4 targeted queries (stats, case studies, expert opinions, pain points)
- 04Real-time web search via gpt-4o-mini-search-preview
- 05Random selection of 1 writing framework out of 4 (editorial variety)
- 06Brand Guidelines from Airtable injected into the system prompt (brand voice)
- 07Theme identification (conceptual/human/strategic/symbolic/reflective)
- 08DALL·E image generation aligned with the style guide
- 09Human approval workflow (Status: Needs Approval)
- 10Auto publishing Mon/Wed/Thu at 7:45am via LinkedIn API
- 11Auto marking Status: Posted + datePosted