What it does (and when to use it)
Use the Agent when you want to go from idea to output fast, without manually picking tools or writing perfect prompts.
The Agent can help with common workflows like:
- Creating new images from a description
- Editing an existing image (add/remove/change objects)
- Generating video from a script, or animating an image into a clip
- Keeping content consistent by using your saved Brand Identity and Personas
- Running saved pipelines to batch-process entire collections
- Scraping product pages and generating content from what it finds
- Executing custom code for pixel-level image manipulation
How to work with the Agent
- Start with a plain-English request (what you want and where you will use it)
- Add context that matters: audience, platform, aspect ratio, tone, and style
- If you have references, upload them (product photo, background, logo, moodboard)
- Ask for iterations: “make it brighter”, “more premium”, “less busy”, “try 3 options”
If you generate for a business, set up Brand Identity and a couple of Personas first. The Agent can use them to keep outputs consistent.
Prompt patterns that work well
- “Create a product hero image for a Shopify listing. Minimal white studio, soft shadow, 1:1.”
- “Remove the background and replace it with a warm kitchen scene. Keep the product size the same.”
- “Use my Brand Identity. Make 5 variations that all look like the same campaign.”
- “Pick a Persona and make a 10 second UGC ad. Script: ‘…’. Add captions.”
- “Animate this image into a short clip with subtle camera movement.”
Credits and limits
Conversations with the Agent are free. When the Agent runs tools like generation, background removal, or upscale, it uses credits from your account.
For programmatic access to all these capabilities, see the Agents & CLI documentation. 
