Generate AI Viral Videos with Seedance and Upload to TikTok, YouTube & Instagram

Generate AI Viral Videos with Seedance and Upload to TikTok, YouTube & Instagram

The JSON you provided is a workflow configuration for an n8n automation pipeline designed to generate AI-powered videos using tools like Wavespeed AI and Fal AI, and then publish them to social media platforms like TikTok, YouTube, Instagram, and others via Blotato. This is an example of an AI agent template for automated video content creation, specifically tailored for generating satisfying, ASMR-style videos involving a sharp blade cutting through materials. Below, I’ll break down the template, its functionality, and how it serves as an AI agent template, keeping it concise and relevant to your query about AI agent templates.

Overview of the Template

  • Purpose: Automates the creation of short, viral-ready videos by generating creative ideas, producing video clips and sounds, stitching them together, and publishing to multiple social media platforms.
  • Platform: Built in n8n, a workflow automation tool, using nodes for AI processing, API calls, and data management.
  • Key Steps:
    1. Generate a creative video idea (e.g., a material being sliced).
    2. Create detailed video prompts for cinematic scenes.
    3. Generate video clips using Wavespeed AI.
    4. Generate ASMR audio using Fal AI.
    5. Stitch clips into a final video.
    6. Upload and publish to social media via Blotato.

Key Components of the Template

This template uses n8n nodes to orchestrate an AI agent workflow. Here’s a breakdown of the main nodes and their roles:

  1. Trigger (Start Daily Content Generation):
    • Type: n8n-nodes-base.scheduleTrigger
    • Initiates the workflow on a schedule (e.g., daily).
    • Acts as the entry point for automated content generation.
  2. AI Agent: Generate Creative Video Idea:
    • Type: @n8n/n8n-nodes-langchain.agent
    • Uses an LLM (GPT-4.1 via OpenAI) to generate a creative concept for a video, following a structured format (e.g., “(Color) (Material) shaped like a (random object)”).
    • Output: A JSON array with fields like Idea, Caption, Environment, Sound, and Status.
    • Integrates a “Think” tool to refine the idea for originality and appeal.
  3. AI Agent: Generate Detailed Video Prompts:
    • Type: @n8n/n8n-nodes-langchain.agent
    • Generates detailed cinematic prompts for three video scenes, focusing on macro-level visuals of a blade cutting through a material.
    • Input: Uses the idea, environment, and sound from the previous step.
    • Output: Structured JSON with detailed scene descriptions (1,000–2,000 characters each).
  4. Video and Audio Generation:
    • Generate Video Clips (Wavespeed AI):
      • Type: n8n-nodes-base.httpRequest
      • Sends scene descriptions to Wavespeed AI’s API to generate 10-second video clips in 9:16 aspect ratio.
      • Waits for clip generation (240 seconds) and retrieves results.
    • Generate ASMR Sound (Fal AI):
      • Type: n8n-nodes-base.httpRequest
      • Generates ASMR audio based on the sound prompt, synced with video clips.
      • Waits for sound generation (60 seconds) and retrieves the output.
  5. Video Stitching:
    • Merge Clips into Final Video (Fal AI):
      • Type: n8n-nodes-base.httpRequest
      • Combines video clips into a single video using Fal AI’s FFmpeg API, with timestamps for seamless transitions.
      • Waits for rendering (60 seconds) and retrieves the final video URL.
  6. Data Management:
    • Save Idea & Metadata to Google Sheets:
      • Type: n8n-nodes-base.googleSheets
      • Stores the idea, caption, environment, sound, and production status in a Google Sheet for tracking.
    • URL Final Video:
      • Updates the Google Sheet with the final video URL and marks production as “done.”
    • Update Production:
      • Marks the video as “Publish” in the Google Sheet after social media upload.
  7. Social Media Publishing:
    • Upload Video to Blotato:
      • Type: n8n-nodes-base.httpRequest
      • Uploads the final video to Blotato’s backend for distribution.
    • Publish to Social Media:
      • Nodes for Instagram, YouTube, TikTok, Facebook, Threads, Twitter, LinkedIn, Bluesky, and Pinterest.
      • Type: n8n-nodes-base.httpRequest
      • Each node sends the video and description to the respective platform via Blotato’s API, using assigned social media account IDs.

How It Fits as an AI Agent Template

  • Modularity: The workflow is broken into reusable steps (idea generation, video/audio creation, publishing) that can be adapted for other content types by modifying prompts or APIs.
  • Automation: The schedule trigger and API integrations enable end-to-end automation, from idea to social media post.
  • AI Integration: Leverages LLMs (GPT-4.1) for creative tasks and external AI services (Wavespeed, Fal AI) for media generation.
  • Scalability: Can handle multiple videos by iterating over scenes or ideas, with batch processing configured in API calls.
  • Customization: Prompts, social media targets, and video parameters (e.g., duration, aspect ratio) can be tweaked for different use cases.

How to Use This Template

  1. Set Up n8n:
    • Install n8n (self-hosted or cloud) and import the JSON workflow.
    • Configure credentials for OpenAI, Google Sheets, Wavespeed AI, Fal AI, and Blotato APIs (replace YOUR_API_HERE with actual keys).
  2. Customize Inputs:
    • Modify the prompt in “AI Agent: Generate Creative Video Idea” for different themes (e.g., change from slicing materials to other ASMR concepts).
    • Adjust video duration, aspect ratio, or social media platforms in the respective nodes.
  3. Run the Workflow:
    • Activate the schedule trigger or run manually to test.
    • Monitor Google Sheets for idea and video metadata.
    • Check social media platforms for published videos.
  4. Extend the Template:
    • Add error handling nodes (e.g., n8n-nodes-base.if for failed API calls).
    • Integrate additional AI tools (e.g., Hugging Face for alternative video generation).
    • Expand to other platforms by adding new HTTP request nodes.

Example Output

Based on the template’s structure, a sample output might be:

  • Idea: “Emerald quartz shaped like a teacup.”
  • Environment: “Polished studio table, soft lighting, reflective surface.”
  • Sound: “Crisp slicing with faint crystalline chimes.”
  • Scenes:
    • Scene 1: “Macro shot: a gleaming blade presses into translucent quartz, micro-cracks forming.”
    • Scene 2: “Close-up: quartz splits smoothly, revealing a sparkling inner texture.”
    • Scene 3: “Over-the-blade shot: fine quartz dust scatters on the reflective table.”
  • Final Video: A 30-second clip combining the three scenes, with ASMR audio, posted to TikTok, YouTube, etc.
Scroll to Top