Automated Faceless Video Creation Pipeline using AI - n8n Workflow

• YouTube Shorts and TikTok creators seeking to automate daily content production.


• Marketing agencies needing rapid, scalable generation of faceless video ads.


• Automation enthusiasts looking for advanced, multi-step n8n templates involving generative AI.


• Solopreneurs aiming to scale content creation without manual editing.

This comprehensive n8n workflow serves as a 'Faceless Video Machine,' automating the entire video production lifecycle from conceptual scripting to final render. It orchestrates several cutting-edge AI services—including Google Gemini (via OpenRouter) for content generation, ElevenLabs for high-quality voiceovers, OpenAI Whisper for transcription and timing, Leonardo AI for scene visualization and video clips, and Shotstack for final video assembly and editing.

By leveraging this advanced n8n workflow, users can bypass tedious manual steps, ensuring content is produced efficiently, perfectly synchronized to the audio track, and ready for immediate publishing. This specific n8n workflow demonstrates the immense power of integrating generative models within an automation platform, ensuring a seamless flow initiated by a simple manual n8n trigger.

The process is initiated by a manual n8n trigger and flows through seven key phases:

1. Input: The manual n8n trigger starts the process. The user defines the video topic within the initial Fields - Set Idea n8n node.


2. Script Generation: The topic is sent to a Langchain n8n node utilizing Google Gemini (via OpenRouter) to write a concise, 60-second video script. This script is then reformatted for optimal voice synthesis.


3. Audio Generation & Transcription: The formatted script is sent to the ElevenLabs API via an HTTP Request n8n node, generating a high-quality voiceover. Simultaneously, this audio is uploaded to Google Drive and sent to the OpenAI Whisper API via another n8n node for detailed transcription, which includes precise word timings.


4. Timestamped Prompt Generation: The original script and the Whisper transcription are merged. This data is fed into a Gemini Langchain n8n node, which uses the timing data to generate descriptive image prompts, each precisely timed to last 2-4 seconds. The output is strictly enforced using an n8n structured output parser.


5. Image Generation: The resulting list of timestamped prompts is split out using an n8n node. Each prompt is then sent to the Leonardo AI API (using an HTTP Request n8n node) to generate visual assets. A 30-second wait n8n node ensures rendering completion.


6. Images to Video Conversion: The generated images are retrieved and sent back to Leonardo AI to be converted into short motion video clips. A 5-minute wait n8n node handles the rendering time. All clips are then downloaded and aggregated into a unified structure, preparing them for the final edit.


7. Final Editing and Download: The aggregated video clips and the public audio link (from Google Drive) are submitted to the Shotstack API (HTTP Request n8n node) for final assembly based on the precise timestamps. After a 1-minute wait n8n node, the workflow retrieves the final polished video for download.

To deploy this powerful n8n workflow, follow these steps:

1. Download: Save the provided n8n workflow JSON code.


2. Import: Log into your n8n instance and use the 'Import from JSON' function to load the n8n template.


3. Credentials Setup: This n8n workflow requires credentials for multiple services. You must configure:

1. Initial Configuration: Locate the Fields - Set Idea n8n node and input your desired video topic string.


2. Execution: Click the 'Test workflow' button, which acts as the manual n8n trigger, or activate and execute the entire n8n workflow.

This complex n8n workflow utilizes numerous specialized n8n nodes:

When clicking ‘Test workflow’ (n8n trigger): The manual trigger n8n node that initiates the entire automation sequence.
60 Second Script Writer (Langchain ChainLLM): This n8n node uses the input idea and connects to the Google Gemini model (via OpenRouter) to produce a creative, engaging video script.
Generate Voice (HTTP Request n8n node - ElevenLabs): Takes the cleaned script text and uses the ElevenLabs API to generate a binary audio file.
Transcribe Audio with OpenAI Whisper (HTTP Request n8n node): Sends the generated audio file to OpenAI's Whisper model to obtain a verbose JSON response containing precise word and segment timestamps, crucial for video synchronization.
Merge (n8n node): Combines the output from the voiceover setup (public URL) and the Whisper transcription data.
Generate Image Prompts (Langchain ChainLLM): This advanced n8n node uses Gemini/Claude to generate a timed sequence of descriptive image prompts based on the transcript's word timings.
Structured Output Parser (Langchain n8n node): This n8n node ensures the output from the LLM is perfectly formatted into a JSON array, making it reliable for subsequent steps.
Generate Images (HTTP Request n8n node - Leonardo AI): Takes the timed prompt and requests a static image from Leonardo AI. This n8n node is executed for every scene.
Generate Videos/Scenes (HTTP Request n8n node - Leonardo AI): Takes the static image ID and requests a motion video clip using Leonardo's Motion SVD model.
Aggregate (n8n node): Gathers all individual downloaded video clips and associated metadata into a single array structure required for the final editing step.

• Edit with Shotstack (HTTP Request n8n node): Submits the combined video clips, along with the Google Drive public audio link and timing data, to the Shotstack API for final high-quality rendering, completing the n8n workflow.