Automates the end-to-end workflow—from receiving a product photo and vision prompt to delivering an AI-generated image and a marketing video via email.
From a submitted photo and vision prompt, the agent converts the prompt into a production-ready instruction and generates a new AI-generated product image. It uploads the image to ImgBB to obtain a public URL and passes the URL to RunwayML to render a product marketing video. The agent monitors the rendering process and emails the customer with links to both assets, while backing up the original photo to Google Drive.
Ingests a reference image and vision prompt, then outputs a new AI-generated image and a marketing video, delivered by email.
Save the original photo to Google Drive as a backup.
Rewrite the user’s vision prompt into a professional image-generation prompt.
Generate a new AI-generated product image using OpenAI’s image editor.
Upload the generated image to ImgBB to obtain a public URL.
Render a product marketing video in RunwayML.
Email the user with links to the image and video via Gmail.
This AI agent eliminates manual image and video production, delivering consistent visuals quickly. It adds reliable backups and automated delivery, reducing risk and delays.
Three-step workflow that non-technical users can follow.
When a user submits the form, the agent saves the original photo to Google Drive as a backup and records the metadata.
GPT-4.1 rewrites the vision prompt into a production-ready image-generation prompt and uses gpt-image-1 to generate the new image.
Upload the generated image to ImgBB for a public URL, render the marketing video in RunwayML, poll every 30 seconds for completion, then email the user with links via Gmail.
One realistic scenario that shows timing and outcomes.
A retailer uploads a product photo and vision prompt. Within about 12 minutes, the agent delivers a brand-new AI-generated product image and a 15-second marketing video, and emails the retailer with links to both assets and a Google Drive backup of the original photo.
Each role gains faster access to usable visuals.
Needs rapid, on-brand visuals for campaigns.
Updates product listings with fresh images and videos.
Prototyping new visuals without design briefs.
Maintains consistency across product visuals.
Requires short-form videos for platforms.
Delivers complete assets to clients quickly.
Key tools used inside the AI agent workflow.
Rewrite prompts and perform image editing (gpt-image-1) inside the agent.
Back up the original photo automatically.
Host the generated image and provide a public URL.
Render the marketing video from the generated image.
Deliver assets to the user via email.
Practical scenarios where this AI agent adds value.
Common questions about how this AI agent works and its outputs.
A user must submit an existing product photo, a name, a vision prompt, and the user’s email. The agent then processes these inputs through a secure flow that backs up the original photo, rewrites the prompt, generates a new image, creates a video, and delivers the assets by email.
The agent outputs an AI-generated product image and a marketing video. The image is hosted on ImgBB and the video is delivered as a shareable asset. Both are linked in the final email sent to the user.
Video rendering can take several minutes depending on length and complexity. The system polls RunwayML every 30 seconds and continues until the video is ready, after which the email is sent. In most cases the process completes in a short session, but longer videos may extend total time.
The workflow uses API keys and OAuth2 credentials for OpenAI, Google Drive, ImgBB, RunwayML, and Gmail. Credentials are stored securely and accessed only by the agent during processing. Access is scoped to the assets being created and delivered.
Yes. The vision prompt guides image and video style, and you can adjust the input prompt to target desired aesthetics. The agent uses RunwayML capabilities to align output with the provided image and messaging, and you can iterate on prompts between submissions.
The agent requires image-generation access via the OpenAI account. If access is temporarily unavailable, the workflow can fall back to alternative prompts or pause until access is restored. In all cases, the original photo backup remains intact, and the notification email will explain any delays.
Only the backup copy stored in Google Drive is kept for archival purposes. The generation process uses the reference image to create a new product image, but the original file remains unmodified in Drive. All assets delivered to the user are links to generated content, not the original photo.
Automates the end-to-end workflow—from receiving a product photo and vision prompt to delivering an AI-generated image and a marketing video via email.
