The studio, inside Blender.
Describe a scene and watch it build. Make pictures from your viewport, video from your scene camera, 3D models from a sketch, and drive the camera with your phone. Every take lands in your Vault; every change is one undo step.
Seven tabs. One prompt box.
Describe a scene. Watch it build.
The chat in the viewport. It looks at your scene, adds the objects, lights them, sets the camera and takes a look, in plain words you can follow. Stop any time; what it did stays.
A texture from your viewport.
Capture the viewport or the last render as a reference, describe the material, and the picture lands as a texture on the selected object, or as a plane.
Video from your scene camera.
Blender renders the frames; the model turns them into the shot. Match timeline sets the length to your range so nothing is cut.
A model from a sketch.
From one picture, textured or plain. The .glb imports where you want it.
Drive the camera with your phone.
Scan the code, hold your phone, walk the room. Rotation and movement go straight to the scene camera; Record keeps the take on the timeline.
Three steps, two minutes.
Claude, ChatGPT, Codex or Cursor can drive Blender.
Turn the Plug on in the MCP tab, connect the app once, and ask it for a scene. Every command it sends shows in Blender’s Activity list; every change is one undo step.
The connect screens are the one place the word MCP appears.Before you install.
The add-on is free. Free mode uses the free chat models and the local tools at no cost. Other models use credits from your Motionbar plan, shown in ⚡ on every button before anything is made.