Create Faceless YouTube Videos on Autopilot with AI + n8n (steal this)

Key Concepts

• Faceless video generation
• AI automation
• Google Sheets integration
• Webhooks
• AI script generation
• Image generation (OpenAI)
• Video creation (RunwayML)
• Audio generation (11 Labs)
• Video merging (Creatomat)
• Captioning (Replicate)
• Social media publishing

Faceless Video Generator Overview

The video demonstrates how to build an automated system to generate and publish faceless videos across various social media platforms like YouTube, Instagram, TikTok, LinkedIn, and X. The system leverages AI and automation tools to create business-related content tailored to specific audiences. The goal is to increase social media presence, generate leads, and ultimately drive revenue.

Workflow Breakdown

The workflow is divided into two main sections:

1. Content Generation: This section focuses on creating the video script, scenes, and other necessary information.
2. Video Creation: This section takes the generated content and assembles it into a final video.

1. Content Generation

• Trigger: The workflow is triggered by updates in a Google Sheet. Instead of using the built-in Google Sheets triggers in n8n, the video recommends using a webhook for more control and reliability.
• Google Sheets Integration:
  • A Google Apps Script is used to send data from the Google Sheet to the n8n webhook.
  • The script is provided and requires the user to copy and paste it into the Google Apps Script editor.
  • The webhook URL from n8n needs to be inserted into the script.
  • A trigger is created in Google Apps Script to send data to the webhook whenever the sheet is edited ("onEdit" trigger).
  • The script requires authorization to access the Google Sheet.
• Data Filtering: A filter in n8n ensures that the workflow only proceeds when column A (the title column) is updated and the updated cell contains text. This prevents the workflow from running on irrelevant changes or empty cells.
• AI Script Generation:
  • An AI agent (using Anthropic's chat model) generates the video script.
  • The script is broken down into four scenes, each designed to be approximately 5 seconds long (213-223 characters including spaces and punctuations).
  • The AI agent is instructed to generate a script that fits within the character limit for each scene.
  • A calculator node ensures that the generated script falls within the specified character range. If the script is too short or too long, the AI agent regenerates it until it meets the criteria.
  • The AI agent also generates a video style and a caption for the video.
  • The generated script, scenes, style, and caption are written back into the Google Sheet.

2. Video Creation

• Trigger: The video creation process is triggered when the value in column I (video status) is set to "to-do".
• Filter: A filter ensures that the workflow only proceeds when column I is updated and the new value is "to-do".
• Scene List Creation: A code node creates a list of the four scenes generated in the previous step.
• Image Prompt Generation:
  • The workflow iterates through each scene and generates an image prompt using an AI agent.
  • The AI agent is instructed to create a clear and compelling graphic prompt suitable for a text-to-image generator, focusing on a marketing style graphic rather than a literal scene.
• Image Generation (OpenAI):
  • The generated image prompt is used to create an image using OpenAI's new image generation API.
  • The video emphasizes the improved quality of the new OpenAI image generation compared to previous versions.
  • The user needs to verify their identity with OpenAI to use the image generation API.
  • An API key is required to access the OpenAI API.
  • The data type of the image is converted to binary data for use in n8n.
• Image Storage: The generated images are uploaded to Google Drive for storage.
• RunwayML Integration:
  • The workflow uses RunwayML to convert the static images into short videos.
  • The Google Drive URLs of the images are formatted correctly for use with the RunwayML API.
  • A loop is implemented with a 20-second delay between each image-to-video conversion to avoid potential rate limiting or server errors with the RunwayML API.
  • The video emphasizes the importance of signing up for the RunwayML dev portal and obtaining an API key.
  • Using RunwayML requires credits and can be expensive.
• Video Status Update: The video status in the Google Sheet is updated to "created".
• Background Music Generation (11 Labs):
  • A sound prompt is generated using an AI agent, requesting soothing, inspirational background music without voices.
  • 11 Labs is used to generate the background music using its sound generation API.
  • A paid 11 Labs account (at least the $5 plan) is likely required.
  • The generated background music is uploaded to Google Drive.
• Text-to-Speech (11 Labs):
  • 11 Labs is used to generate a voiceover from the video script.
  • The user can choose from various AI voices or even use a synthesized version of their own voice.
  • The generated voiceover is uploaded to Google Drive.
• Video Merging (Creatomat):
  • Creatomat is used to merge the video clips, background music, and voiceover into a single video.
  • Two merge nodes are used to combine the video clips and audio files.
  • A code node extracts the URLs of the video clips into a list.
  • A Creatomat template is used to define the video layout and composition.
  • The video provides a pre-made Creatomat template in the description for users to copy and paste.
  • The user needs to create a Creatomat account, obtain an API key, and update the template ID in the n8n workflow.
  • Using Creatomat requires a paid subscription.
• Captioning (Replicate):
  • Replicate is used to add captions to the final video.
  • The user needs to create a Replicate account, obtain an API token, and configure the n8n workflow.
  • The video is passed to the Replicate API, which automatically overlays the captions on top of the video.
• Video Publishing (YouTube):
  • The captioned video is uploaded to YouTube as a short.
  • The video can be published immediately or scheduled for later.
• Final Video Storage: The final video is uploaded to Google Drive, and the link is added back to the Google Sheet.

Tools and Services Used

• n8n: The automation platform used to orchestrate the entire workflow.
• Google Sheets: Used as the data source and to track the video creation process.
• Google Apps Script: Used to connect Google Sheets to n8n via webhooks.
• Anthropic: AI model used for script generation.
• OpenAI: Used for image generation (DALL-E 3).
• RunwayML: Used for converting images into short videos.
• 11 Labs: Used for generating background music and text-to-speech voiceovers.
• Creatomat: Used for merging video clips, background music, and voiceovers into a single video.
• Replicate: Used for adding captions to the final video.
• YouTube: Used as the primary social media platform for publishing the generated videos.

Costs and Considerations

• The workflow involves several paid services, including RunwayML, 11 Labs, Creatomat, and Replicate.
• Using the OpenAI image generation API requires identity verification.
• RunwayML can be expensive due to its credit-based pricing.
• The video highlights potential issues with the RunwayML API, such as internal server errors, and suggests implementing a loop with a delay to mitigate these issues.
• The video suggests using make.com for publishing to other social media platforms due to its better native integrations.

Notable Quotes

• "This is a fantastic starting point for you to take an automation like this and build build on top of it to create something that really really works for your particular industry and use case scenario."
• "I don't want this to be like a stumbling block because there are a couple lines of code here but you could certainly get around this by just asking chat GBT the right questions."
• "You are freaking awesome don't forget that."

Technical Terms and Concepts

• Webhook: A mechanism for one application to send real-time information to another application whenever a specific event occurs.
• API (Application Programming Interface): A set of rules and specifications that software programs can follow to communicate with each other.
• JSON (JavaScript Object Notation): A lightweight data-interchange format that is easy for humans to read and write and easy for machines to parse and generate.
• Data Structures: Ways of organizing and storing data in a computer so that it can be used efficiently.
• Rate Limiting: A mechanism to limit the number of requests a user can make to an API within a given time period.
• Binary Data: Data that is represented in the form of bits (0s and 1s).

Synthesis/Conclusion

The video provides a detailed walkthrough of building a sophisticated faceless video generator using AI and automation tools. It covers the entire process from content generation to video publishing, highlighting the various services and APIs involved. While the workflow requires some technical knowledge and involves several paid services, it offers a powerful solution for automating video creation and distribution across social media platforms. The video emphasizes the importance of customization and encourages viewers to adapt the workflow to their specific needs and industries. The creator also promotes his school community as a resource for those seeking to further their knowledge of AI automation and business scaling.