rigami.
All stacks
/// AI · Frontier LLM

Gemini 3 + Nano Banana for vision and multimodal.

Gemini 3 for large-context multimodal reasoning and Gemini Nano Banana for image generation — wired into product surfaces with the right cost ceilings.

/// What we ship with Google Gemini

How we use it
in production.

  • Gemini 3 Flash / Pro for chat and reasoning
  • Nano Banana for image generation
  • Multimodal input — text, image, audio, video
  • Grounded search and retrieval
  • Streaming and structured outputs
  • Cost-aware model routing