Content Autopilot
Finished content from infrastructure you own.
Content Autopilot is a self-hosted n8n pipeline that replaces Blotato with local inference. Local models run through Ollama, and the workflow routes output into three formats: text posts, carousels and short-form video.
Conceptual illustration, not a product screenshot.
- Inference runs locally on Ollama, with Groq as an optional cloud fallback for RAM-constrained batch processing.
- Short-form video uses third-party services: ElevenLabs text-to-speech and Pexels stock footage. Carousels render through a local Puppeteer service and video assembles with FFmpeg.
- This reduces tool subscriptions but is not a zero-external-dependency system.
A GrowthBoxx-native n8n pipeline using local Ollama models (Qwen3, Llama 3.2, Phi-4 Mini) to produce text posts, carousels and short-form video.
Qwen3 for briefs
Structures the content brief.
Llama 3.2 for copy
Writes platform-native copy.
Phi-4 Mini for polish
Handles formatting and finishing.
Three output paths
Text posts logged to a Markdash-compatible Google Sheet, carousels, and short-form video.
Optional Groq fallback
Cloud inference when local RAM limits batch work.
Local rendering services
Puppeteer for carousels and FFmpeg for video assembly.
One brief, three outputs
Illustrative workflow diagram. It shows the sequence of steps, not data or results.
- Owned infrastructure
The pipeline runs on your own or the operator's machines.
- Right model per stage
Different local models handle planning, writing and polish.
- Reporting-ready
Text output logs to a Sheet that Markdash can read.
- 1Brief
Qwen3 structures the content brief.
- 2Write
Llama 3.2 drafts platform-native copy.
- 3Polish
Phi-4 Mini formats and refines.
- 4Route
Text, carousel or video path is chosen.
Offered as a productized monthly retainer with white-labeled output. The client receives finished content; the operator owns the infrastructure.
Keep several channels fed without a content department.
Deliver white-labeled content as a monthly service.
Turn one brief into posts, carousels and video.
What clients say about working with us
General feedback about working with GrowthBoxx. These are not endorsements of Content Autopilot specifically.
“We went from manually sourcing 20 leads a week to running a fully automated pipeline that delivers over 80 qualified prospects daily. The AI integration into our CRM changed how our entire sales team operates — we closed three enterprise deals in the first month we wouldn't have had capacity to pursue before.”
“Our content team was a bottleneck — we were producing four articles a month and spending a fortune doing it. With the Content Autopilot system GrowthBoxx built, we're now publishing daily across three channels. Traffic doubled in six weeks and our cost per piece dropped by more than two-thirds. The quality actually improved.”
“I was sceptical about AI personalisation at scale — I'd seen it done badly. What GrowthBoxx built was different. Every message felt genuinely researched. Our reply rate on cold outreach went from 6% to over 44% in the first campaign. We hit our quarterly revenue target in week seven. Best investment we made this year.”
Honest answers
Is anything sent to the cloud?+
Core inference is local through Ollama. Groq is optional, and video uses ElevenLabs and Pexels.
Does it replace human review?+
We recommend reviewing output before publishing. This is a pipeline, not a guarantee of results.
Who owns the infrastructure?+
The operator owns it. The client receives the finished content.
Talk through Content Autopilot with us.
Book a call to discuss whether it fits your team, how it would be set up, and where its limits are.
Book a call about Content Autopilot
