AI-Powered Short Video Ad Generation from Image Input - n8n Workflow

Digital Marketers: Rapidly deploy tailored video ads across social platforms without manual video editing.
E-commerce Owners: Generate dynamic product showcases directly from static product photos.
Content Creators: Users looking for robust n8n templates to handle complex, multi-step generative AI tasks.
Automation Specialists: Anyone seeking an advanced n8n workflow example incorporating multiple asynchronous API calls (polling) and video sequencing.

Creating high-quality, engaging video ads for social platforms like TikTok or Instagram Reels is time-consuming and resource-intensive. This comprehensive n8n workflow solves this by fully automating the creative process. Starting from a single product image and a short description, the automation leverages the power of the Gemini 2.5 Pro model to construct a detailed four-part visual storyboard, including specific prompts for image edits and subsequent animations.

The core of this n8n automation then executes these steps: 1) It generates four distinct, AI-edited images using Gemini 2.5 Flash. 2) It transforms each edited image into a dynamic video segment using the FAL WAN Image-to-Video model. 3) It sequences these segments using FAL FFmpeg and adds bespoke audio. The final output is ready to be posted directly to multiple social platforms via the Upload Post n8n node. This entire n8n workflow is designed for high efficiency and scalability.

This complex n8n workflow operates in several distinct phases, ensuring smooth execution of generative AI tasks:

1. Trigger & Input: The process starts with the Photo Upload Form n8n trigger node, where the user uploads an original product photo and provides a text description.


2. Storyboard Generation: The description is fed to the Storyboard Agent (powered by Google Gemini Chat Model and Structured Output Parser), which meticulously defines four frame prompts (prompt1–prompt4) and corresponding image-to-video animation prompts (i2vprompt1–i2vprompt4) in a strict JSON format.


3. Image Processing: The original photo is uploaded to imgbb. Four parallel branches execute, each using an HTTP Request n8n node to call the Gemini 2.5 Flash API. This generates four unique, edited versions of the original image based on the generated storyboard prompts. These four edited images are then uploaded to imgbb to obtain stable public URLs.


4. Video Animation (Parallel Polling): The four image URLs are passed to four parallel branches. Each branch calls the FAL WAN Image-to-Video API via an HTTP Request n8n node to start the asynchronous video job. A polling loop (Wait -> FAL WAN i2v (status) X -> Animation Completed? X If n8n node) ensures the n8n workflow waits until the video segment is generated before proceeding to fetch the final video URL.


5. Composition & Sound: Once all four video segments are retrieved, the Sequence Video (FFmpeg) n8n node compiles them into a single 20-second video. A subsequent parallel process generates ambient audio (Create Sounds) based on the storyboard's sound prompt.


6. Final Output: The finished video (with audio track) is downloaded and automatically prepared for multi-platform posting using the dedicated Upload Post n8n node. This completes the sophisticated n8n workflow.

To deploy this powerful n8n workflow template, follow these steps:

1. Import: Copy the provided JSON data and import it directly into your n8n instance as a new n8n workflow.


2. Credentials Setup: This n8n workflow requires several external credentials:

1. API Key Configuration: Locate the Set APIs Vars n8n node and update the value for imgbbapikey with your actual imgbb token.


2. Activation: Ensure the Photo Upload Form n8n trigger node is correctly activated and the webhook URL is accessible. Run the n8n workflow to test the automation from the form interface.

This n8n workflow relies on a complex arrangement of standard and specialized n8n nodes:

Photo Upload Form (n8n trigger): The starting point, accepting an image and a 'Product description' text field.
Storyboard Agent (Langchain Agent n8n node): Utilizes the Google Gemini Chat Model and a Structured Output Parser to convert the product description into structured, executable prompts for 4 image edits and 4 animation sequences.
Set APIs Vars (Set n8n node): Stores configuration variables, most critically the imgbbapikey and image generation settings.
HTTP Request (n8n node): Heavily used throughout the n8n workflow. Key uses include:
Uploading original and edited images to imgbb.
Calling Gemini 2.5 Flash for image editing (Gemini 2.5 Flash - Generate Image X).
Initiating and polling the FAL WAN Image-to-Video service (FAL WAN i2v (Queue) X, FAL WAN i2v (status) X, and FAL WAN i2v (Result) X).
Calling FAL FFmpeg for video sequencing (Sequence Video (FFmpeg)).
Calling FAL MMAudio for sound generation (Create Sounds).
Wait (n8n node): Crucial for managing asynchronous tasks (like video animation and composition) by creating polling delays until the results are ready.
If (n8n node): Used in the polling loops (Animation Completed? X) to check if the video generation status is 'COMPLETED' before proceeding.
Merge / Aggregate (n8n node): Essential for combining the results of the parallel image editing and video animation branches before sequencing.

• Upload Post (n8n node): The final action, downloading the video data and uploading it to social platforms (TikTok, Instagram, YouTube) defined in the n8n template.