Dev Tools · 1h ago
Kiro AI Skill Uses Gemini's Stateful API for Iterative Image Editing
A new open-source skill for the Kiro coding agent wraps Google's gemini-3.1-flash-lite-image model, enabling stateful image generation and editing via the Interactions API. Users can iteratively refine images without re-prompting the entire scene, with sub-2-second generations and text rendering in 25+ languages. The skill packages the model as an MCP server and a Kiro skill, demonstrating practical use of persistent visual context.
Meridian48 take
The stateful approach solves a real pain point in AI image generation, but its reliance on Google's proprietary API and server-side context storage raises questions about vendor lock-in and data privacy.
Read the full reporting
Teaching Kiro to Paint: A Stateful Image-Editing Skill Built on Gemini's Interactions API and MCP →
DEV Community
ai-image-generationdeveloper-tools