Conversational image segmentation with Gemini 2.5- Google Developers Blog

Introduces conversational image segmentation with Gemini 2.5, enabling users to segment objects in images using natural language queries that describe relationships, conditional logic, abstract concepts, in-image text, and multi-lingual labels. This feature enhances interactive media editing, intelligent safety and compliance monitoring, and insurance damage assessment. The post highlights the simplified developer experience with a single API for advanced vision applications and provides links to demos and documentation for building with the Gemini API.