Native Image Generation API Integration Using AgenticFlow Tokens

I’d like to propose a native image generation API integration directly within AgenticFlow using AgenticFlow tokens, rather than relying solely on PixelML or BYO keys. While I understand this feature might be resource-intensive due to GPU costs, some workflows, especially those involving newer models like DeepSeek R1, include image outputs as part of the reasoning or user experience.

It would be great if the AgenticFlow team could explore a more sustainable and community-supported solution:

• A lightweight or shared GPU backend optimized for low-volume image generation.

• An optional token-tiered system where image generation is available but capped to encourage prototyping.

• A community feedback loop or waitlist to gauge actual demand before scaling infrastructure.

Additionally, if cost remains a major concern, perhaps consider charging image generation requests using AgenticFlow tokens at a higher token rate. This way, users who only need image generation for proof-of-concept (PoC) or occasional use can still access it without being forced to purchase third-party tokens up front.

Thanks for considering this!