Proposal: Extend Cerebras Integration with Image and Multimedia Support

[tallesdantas]

Cerebras has proven to be, in my experience, one of the most efficient and precise solutions for code-related tasks — both in terms of speed and accuracy. However, its current capabilities could be significantly enhanced with the addition of image and other multimedia support.

A strong reference for this type of functionality can be found in the work done by the Augment project. They have already implemented similar features, and their repository may provide useful insights for extending Cerebras in RooCode:

👉 Augment GitHub Repository

I believe implementing these capabilities would make RooCode even more powerful and versatile, especially for projects that combine code reasoning with visual or multimodal inputs.

I’d be genuinely excited to see this improvement, as I rely on Cerebras heavily and appreciate its current strengths.