Skip to content

Feature: Support image output from Grok (Aurora image generation) #586

@gucasbrg

Description

@gucasbrg

Feature Request

When using grok ask to request image generation, Grok generates images on the web page but the CLI only returns the text response. The generated images are lost.

Current Behavior

opencli grok ask "draw a cute cat"
# Returns only text: "好的!给你画几只超级可爱的猫咪~ 😺"
# Images generated on grok.com are not captured

Expected Behavior

opencli grok ask "draw a cute cat" --output ./images
# Returns text response + downloads generated images
# Or returns image URLs

Suggested Output

Response Images
好的!给你画几只... /tmp/grok-img-1.png, /tmp/grok-img-2.png

Technical Notes

Grok uses Aurora model for image generation. The images are rendered as <img> elements on the page. The CDP/browser adapter should be able to:

  1. Wait for image generation to complete
  2. Extract image URLs or base64 data from the page
  3. Download and save images locally or return URLs

Use Case

AI agents need to generate and retrieve images through Grok's free image generation capability without manually opening a browser.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions