Google Veo 3 Fast AI Video Generation with Replicate Polling - n8n Workflow

Content creators needing programmatic video generation.
Developers integrating cutting-edge AI video models into applications.
n8n users looking for advanced examples of asynchronous job polling mechanisms.
Automation specialists seeking reliable n8n templates for external APIs like Replicate.

Generating high-quality, minute-long videos using advanced models like Google Veo 3 Fast is an asynchronous process—it takes time. Standard HTTP requests in n8n might time out waiting for the result. This robust n8n workflow template solves this challenge by implementing a complete polling loop.

Starting with a simple n8n trigger, the workflow initiates the video creation request via Replicate, then continually checks the status until the video generation job is marked as 'succeeded'. This ensures that even lengthy AI tasks are handled reliably within your n8n infrastructure. This pattern is essential for maximizing the utility of complex AI services within your n8n automation projects.

This powerful n8n workflow begins with the 'On clicking 'execute'' manual n8n trigger, allowing for immediate testing or simple manual invocation.

1. Initialization: The 'Set API Key' n8n node first securely stores your Replicate API key for subsequent HTTP requests.


2. Creation Request: The 'Create Prediction' n8n node sends a POST request to Replicate to start the video generation using the google/veo-3-fast model, providing the desired prompt and configuration.


3. Data Extraction: The 'Extract Prediction ID' code n8n node processes the initial response, extracting the unique job ID and the full prediction URL required for polling.


4. Polling Loop Start: The workflow enters a loop starting with the 'Wait' n8n node, which pauses execution for 2 seconds to prevent excessive API calls while the video is being processed.


5. Status Check: The 'Check Prediction Status' n8n node repeatedly queries the specific prediction URL using the extracted ID and the authorization header.


6. Conditional Branching: The 'Check If Complete' If n8n node evaluates the current status received from Replicate. If the status is 'succeeded', the process exits the loop. If the status is anything else (e.g., 'processing'), the flow loops back to the 'Wait' n8n node to continue polling.


7. Result Processing: Once successful, the 'Process Result' n8n node extracts the final video URL and relevant metadata (such as timestamps and metrics), concluding the n8n workflow execution.

To deploy this video generation n8n workflow template successfully, follow these steps:

1. Import: Copy the provided JSON code and paste it into your n8n instance via the 'New' menu -> 'Import from JSON'.


2. API Key Setup: Locate the 'Set API Key' n8n node. Update the replicateapikey assignment with your actual API key from Replicate.


3. Prompt Configuration: In the 'Create Prediction' HTTP Request n8n node, modify the JSON body to include your desired prompt for video generation. You can also adjust parameters like seed.


4. Activation: Ensure the n8n workflow is active (toggle switch in the top right).


5. Execution: Click the 'Execute Workflow' button on the 'On clicking 'execute'' n8n trigger to run the process and generate your video.

This n8n workflow leverages several specialized n8n nodes to manage the asynchronous API process:

On clicking 'execute' (Manual Trigger n8n trigger): Serves as the starting point for this n8n workflow template, allowing manual initiation of the video generation job.

Set API Key (Set n8n node): Used to inject the Replicate API key securely into the execution context, ensuring it is available for all subsequent HTTP Request n8n nodes.

Create Prediction (HTTP Request n8n node): The crucial step where the POST request is sent to api.replicate.com to initiate the Veo 3 Fast model run. It uses HTTP Header Authentication with the dynamically set API key.

Extract Prediction ID (Code n8n node): A custom n8n node script that parses the initial prediction response JSON to extract the required predictionId and constructs the polling URL for reuse.

Wait (Wait n8n node): Implements a necessary time delay (2 seconds) within the polling loop. This prevents API rate limiting and ensures resource efficiency for the n8n workflow.

Check Prediction Status (HTTP Request n8n node): This n8n node repeatedly polls the Replicate API using the unique predictionUrl stored from the previous step, checking the progress of the video generation.

• Check If Complete (If n8n node): The core control flow n8n node for the polling mechanism. It directs the flow back to the 'Wait' n8n node if the status is not 'succeeded' or forward to processing if complete.