Imagent – agentic image/video/speech generation

Imagent is a unified interface that enables AI agents to generate images, video, and speech as a native step in their workflows, abstracting away differences between various providers and models. It is designed for developers and AI engineers building autonomous agents that need multimodal output capabilities. The project is interesting because it treats media generation as a first-class action for agents, simplifying integration and unlocking more dynamic, interactive agent behaviors.