n8n has released a publicly accessible workflow template that automates the conversion of RSS blog posts into 32-second AI-generated video Shorts. The pipeline relies on Claude Opus 5.5 to generate edit specifications and Shotstack to render the final vertical 9:16 video. This is not a vague concept demo; it is a documented, end-to-end blueprint for teams looking to repurpose existing written content into short-form video without building the architecture from scratch.

The Architecture: Claude Writes, Shotstack Renders

The workflow is structured around three distinct stages: content input, AI-driven edit creation, and video rendering. The RSS feed serves as the source material, feeding blog post text directly into the system. Claude Opus 5.5 then processes this text to write a Shotstack edit specification. Unlike simple summarization tasks, Claude’s output here is a structured set of instructions that dictates voiceover, subtitles, animated text, and music timing. Shotstack receives these instructions and renders the final 32-second video asset. The template is unusually specific about the intended output structure. It references the use of 5 'Nano Banana' assets, two images, and ElevenLabs music set at 120 BPM. The animated text is synchronized to the beat, ensuring the result is a timed, dynamic edit rather than static text over a background. This level of detail demonstrates that the workflow is designed for production-quality short-form content, not just a proof of concept.

Orchestration vs. Editorial Control

The core value of this template lies in orchestration. n8n connects the disparate tools—RSS, LLM, and video renderer—into a single automated flow. However, the automation does not eliminate the need for human oversight. The workflow provides a repeatable format, but it does not decide which blog posts are worth converting. Teams must still curate source material, define brand voice parameters, and determine how the resulting Shorts fit into their broader distribution strategy. For smaller teams, this separation of concerns is critical. The template handles the repetitive coordination between content selection, AI processing, and rendering. This allows editors and marketers to focus on high-level decisions about content quality and audience targeting. The published example serves as a starting point for evaluation, enabling companies to test whether their specific RSS feed and brand style align with the generated output.

Key Takeaways

  • The template uses Claude Opus 5.5 to generate Shotstack edit instructions, not just summaries.
  • Output is strictly defined as a 32-second vertical (9:16) video.
  • The workflow integrates ElevenLabs music (120 BPM) and specific asset types like 'Nano Banana' images.
  • n8n handles the orchestration, but editorial curation of source posts remains a human task.

The Bottom Line

This is a practical blueprint for automating the tedious handoffs in content repurposing, but it’s not a magic bullet for content strategy. If you have a steady publishing rhythm, this saves time; if you don’t have good source material, it just generates bad videos faster.