Enable image input support in Amazon Q CLI for enhanced visual development workflows

[dpmramesh]

Checks

• I have searched github.com/aws/amazon-q-developer-cli/issues and there are no duplicates of my issue
• I have run q doctor in the affected terminal session
• I have run q restart and replicated the issue again

Operating system

macOS 15.3.2 (24D81)

Expected behaviour

Background

Amazon Q CLI currently lacks support for processing image inputs, limiting its ability to assist with visual development workflows.

Problem Statement

Developers and designers cannot leverage Amazon Q CLI's capabilities for image-based tasks, such as:

• UI component generation from screenshots
• Design review assistance with diagrams
• Visual asset analysis and suggestions

Proposed Solution

Add image input support to Amazon Q CLI to enable:

1. Direct image file processing
2. Screenshot analysis
3. Visual context understanding
4. UI component suggestions based on visual inputs

Benefits

• Improved efficiency for UI/UX development workflows
• Enhanced collaboration between design and development teams
• Faster prototyping and iteration cycles
• Better support for visual documentation analysis

Success Criteria

• Amazon Q CLI can accept and process image file inputs
• Support for common image formats (JPG, PNG, etc.)
• Ability to generate relevant code suggestions from visual inputs
• Integration with existing development workflows

Implementation Considerations

• Image processing capabilities in the backend
• Security and privacy considerations for image data
• Performance impact on the CLI
• Integration with existing Amazon Q features

Actual behaviour

Feature not supported

Steps to reproduce

No response

Environment

[q-details]
version = "1.7.2"
hash = "b6bf5f1675dd7395009e8bc65c1145820671f211"
date = "2025-03-21T01:36:16.66764Z (6d ago)"
variant = "full"

[system-info]
os = "macOS 15.3.2 (24D81)"
chip = "Apple M1 Pro"
total-cores = 10
memory = "32.00 GB"

[mschrage]

The API currently supports this. We need to add give a mechanism for users to include images in requests (UX) and then encode it as part of the request.

This could be a /image command and then pass the path to a file? Open to ideas and discussion