How to Convert a Video into AI Prompts (Scene-by-Scene Workflow)

To turn a reference video into reusable AI prompts, break it into scenes by timestamp, then for each scene write a start-frame prompt, an end-frame prompt, and motion directions. The start and end frames should describe the same subject, colors, and lighting so the AI tool can interpolate between them; the motion directions tell the tool how the camera and subject move. This workflow suits anyone who wants to recreate the look of a YouTube video or Short in Sora, Veo, or a similar generator, and it works best when the reference video has clear, stable shots rather than rapid cuts.

What a finished prompt set looks like

A complete scene breakdown has three parts. Using the product-reveal example from TubePrompter:

Start frame prompt

A matte cream ceramic mug on a warm oak table beside a window. Medium close-up at table height, soft daylight from camera left, gentle shadow to the right, muted beige background, the entire mug visible, no text or logos.

End frame prompt

The same cream ceramic mug and oak table, now in a tighter close-up of the handle and rim. Keep the soft left-side daylight, beige background, and warm colors consistent; the mug remains still.

Motion directions

Add a slow camera push toward the handle. Keep the mug fixed in place and end on the close-up. Set the shot duration in your video tool.

The start and end frames share subject, palette, and light direction; only framing changes. That consistency is what lets the tool generate a smooth move instead of a jump.

Step-by-step workflow

1. Split the reference video into scenes

Watch the video and note timestamps where the shot changes — a cut, a new angle, or a new subject. Each block between two timestamps becomes one scene. TubePrompter's own breakdown lists this as "scene timestamps → visual notes → start and end frame prompts," so the timestamp list is the first artifact you produce.

2. Write visual notes per scene

For each scene, record:

• Subject — what is on screen and what it is doing
• Framing — wide, medium, close-up, vertical or horizontal
• Camera — height and angle (e.g. "at table height")
• Lighting — direction and quality (e.g. "soft daylight from camera left")
• Background and palette — e.g. "muted beige background," "restrained natural colors"

3. Draft the start-frame prompt

Describe the scene as a single still. Be specific about framing, light direction, and color. The landscape example shows the level of detail expected: "Wide landscape view from a rocky ridge at sunrise. Dark grass and stones in the foreground, a misty valley in the middle distance, layered blue mountains on the horizon, warm light along the ridge, restrained natural colors."

4. Draft the end-frame prompt

Describe where the shot ends, keeping everything except framing identical. In the landscape example the end frame moves "slightly farther forward so the foreground rocks sit near the lower edge" while preserving "the mountain shapes, pale mist, and warm rim of light."

5. Add motion directions

State the camera move and any subject movement, then set duration in your video tool. Examples from the source material:

• "Add a slow camera push toward the handle. Keep the mug fixed in place and end on the close-up."
• "Move the camera forward slowly with a level horizon. Let the foreground pass out of frame while the distant mountains remain steady."

6. Copy into your video tool

Paste the start prompt, end prompt, and motion directions into Sora, Veo, or your generator of choice, then adjust wording based on the first output.

Common reasons prompts underperform

Tool note

TubePrompter generates scene-by-scene prompts with camera, lighting, and style details from YouTube videos and Shorts, and its site states that signing in gives 5 free analyses per month. Check the pricing page for current plan details before relying on a specific free allowance.