AI Video Studio

Current access

New paid subscriptions are temporarily paused. Existing subscriptions are unaffected.

Core chat, reasoning, synthesis, creative modes, and Image Studio are available on the Free plan. Video Studio, live web search, and persistent memory require Pro access.

Create cinematic video

Describe the scene, movement, camera direction, atmosphere, and visual style. MindGod can generate from text alone or use an uploaded image as the starting point.

Choose a 4–60 second total length in 4-second steps. Text videos include generated audio; reference-photo videos are silent. Landscape 16:9 and portrait 9:16 are aspect ratios, not exact pixel-size promises.

Requests longer than 12 seconds are assembled from independently rendered clips, so visual continuity between clips is not guaranteed.

The studio displays generation progress and provides the completed video for playback and download.

Preview a full-body character before making a video

Enter the scene, face, clothing, setting, and movement together in one description, up to 2,500 characters. An optional reference can be a face-only or cropped photo; a full-body source image is not required.

A full-body request creates a character preview first. Review the face, clothes, proportions, and framing. Choose Regenerate character to make another preview, or Proceed to video to animate the exact approved image. Changing the description, photo, or style clears the old preview.

Each preview or regeneration is a separate image request. Video generation begins only after approval, requires Pro access, and uses the approved frame as its reference. AI invents unspecified details; likeness and exact visual continuity are not guaranteed.

How to animate a reference photo

Use an image you own or have permission to animate. Keep the request focused: describe the subject's movement, camera movement, and background motion rather than packing several unrelated scenes into one prompt.

For example: “A gentle breeze moves the leaves, the subject looks toward the camera, slow push-in.” Reference-photo videos are silent. Start with a short render to evaluate movement before requesting a longer sequence; resemblance and motion can vary and exact identity preservation is not guaranteed.