Visual Regression Testing with Apify and AI Vision Model

This workflow is designed to facilitate visual regression testing by leveraging the capabilities of Apify for screenshot generation and an AI vision model for analysis. The primary goal is to ensure that visual elements of a web application maintain their appearance across different deployments and updates. By generating baseline screenshots and comparing them against the current state, developers can quickly identify unintended changes in the UI. This workflow is particularly beneficial for web developers and QA teams who need to maintain high standards of visual consistency across various devices and browsers.

The workflow initiates by downloading base images from Google Drive, which serve as the reference points for comparison. It then utilizes the Google Gemini Chat Model to analyze the images and provides structured output that details any discrepancies found during the comparison. This structured output includes information about the type of regression, descriptions of the changes, and the previous and current states of the UI elements.

Furthermore, the workflow is equipped with a sticky note feature that outlines the importance of generating base images before initiating the testing process. This ensures that users are aware of the necessity of having an updated set of baseline images before conducting regression tests. Overall, this workflow streamlines the visual testing process, reduces manual efforts, and enhances the reliability of web applications by automating the detection of visual regressions.