Capability
Media generation
Text-to-image and text-to-video via MCP across multiple vendors, with reference images, first-frame drive and inpainting — generate right in the chat.
How it works
Multi-vendor
DashScope, Kling, Seedance, Seedream, OpenAI, Veo.
Image & video
Pick a model in the composer, or ⚡ generate directly.
Controllable
Reference i2i, first-frame i2v, mask inpainting.