AI can generate a beautiful shot and still direct the wrong film.

The face looks right. The camera floats somewhere random. The actor misses the movement.

Another paragraph in the prompt will not always fix this. The model needs better visual instructions.

Those instructions can be a finished image, a storyboard, an existing shot, a depth map or a rough 3D scene. Each one gives you more control over what appears on screen and how it moves.

Here are the five workflows I would learn in order

I am Alex, welcome to ShortCu8 by Innov8.

Lets Dive Deep 🐰

Build a reference board before opening a video model.

Collect the things your film must keep consistent:

Use Pinterest or Cosmos to collect the wider visual direction. When you need a precise emotion or cinematic composition, search FrameThrower for matching frames from real films.

Generate the starting images for your scenes. My current picks are:

Give the board and your scene intention to an LLM. Let it draft the image and video prompts, then correct anything that misses your vision.

This preparation stays the same across all five levels.

Create one finished base image for every scene. Then animate each image separately and edit the clips together.

You control the opening frame. The model still decides much of the motion and camera behaviour.

My current picks are Seedance 2.0 and Gemini Omni.

Instead of generating every scene alone, create one 4x4 storyboard containing 16 connected panels.

Show the start frame, camera changes, action beats and ending. Give the complete board to Seedance or Gemini Omni as one sequence.

This 16-panel Oni workflow shows the method.

For tighter spatial control, create a second version of the board as a depth map. The normal board describes appearance.

The depth-map storyboard describes which objects are near, far and moving through the scene.

Here, an existing video becomes the direction.

Give the model a reference clip plus an image of your new character. It can borrow the shot order, camera movement and performance.

This is how many viral football edits place Messi or Ronaldo inside recognisable film scenes. The movie clip supplies the direction. The character image supplies the replacement.

You can also use EyeCandy to find pans, tracking shots and music-video camera moves. Download a shot, pair it with your character image, and describe what must stay unchanged.

Use Level 3 when you can already point to a shot and say, "I want this camera language with my character."

This level is for copying movement itself: a dance, fight, hand gesture or full-body performance.

Kling Motion Control was the familiar option. Seedance 2.0 and Gemini Omni now give you another route using character and scene references.

Fine movement can disappear when you provide only the original clip. Extract a depth video with fal.ai Depth Anything, then pair it with the new character image.

The depth pass removes the source video's appearance while preserving body position and movement. See the depth-guided Seedance workflow.

Use Level 4 when the motion must match, not merely feel similar.

This gives you the most control.

Build a rough scene in Blender or Unreal Engine. Place basic objects, block the action and animate the camera. Its job is to define position and movement.

Export the preview. Give it to Seedance or Gemini Omni with your character and style references, then ask the model to rebuild the rough geometry in your finished visual style.

These Blender reference examples, camera blockouts and 3D-guided shots show the idea. Here is a similar Unreal Engine and AI workflow.

If you do not know Blender, Blender MCP can help an AI agent create the blockout. You still need to check the camera and staging.

Now go and create something great!

We read every reply - just reply to this email and let us know how we can improve !

#AD1

The best assistants don't multitask their attention across a hundred tools. Neither does Catch. It's an AI agent that focuses on one thing — the admin work you'd rather not touch — and does it exceptionally well.

Scheduling, flights, restaurants, follow-ups, vendors, clients. You hand it over; Catch handles the back-and-forth and comes back with it done.

No context-switching. No dropped balls. Just your admin, quietly cleared — so your focus stays on the work only you can do.

Meet the agent built for admin, and it'll be ready to work before your next meeting.

Get started at catchagent.ai — and give your attention back to what matters.

Start now for free

#AD2

SSP under ASC 606 shouldn't live in spreadsheets. On July 22, Tabs' team shows how native SSP automation runs on real billing data, ties to your Product Catalog, and generates audit-ready documentation automatically.

July 22 · 1:30–2:00 PM EDT · Live + recording

Reserve your seat