Agent
generator
Produces a register for the target context
@cf/openai/gpt-oss-120b
With image input: @cf/meta/llama-4-scout-17b-16e-instruct
Takes your master house style and produces a register for a specific context - denser, darker, more editorial - without letting it stop being yours. Two scores pull against each other: how faithful it stays to the signature, and how well it serves the new context. It loops until both clear, and a legibility failure fails the whole run regardless of either.
The team
Agent
evaluator
Scores fidelity and fit, and gates on the floors
@cf/openai/gpt-oss-120b
With image input: @cf/meta/llama-4-scout-17b-16e-instruct
Tools: contrast_ratio
Agent
optimizer
Fixes floors first, then the lower of the two axes
@cf/openai/gpt-oss-120b
With image input: @cf/meta/llama-4-scout-17b-16e-instruct