Skip to content

[FEATURE] Vision Proxy for non vision model #74

Description

@wdw89

✨ The feature

when pasted an image or mcp return an image for non vision model, can use a vision model like qwen-3.7-plus or minimax m3 to see the image, then return text description for non vision model.

🎯 The problem

add vision capbility to non vision model, now need to use an vision mcp or vision subagent to do this, and not support pasted image.

Metadata

Metadata

Assignees

No one assigned

    Labels

    enhancementNew feature or request

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions