AI Video Prompt Generator: Shot-by-Shot Template

Video goal: [what the viewer should understand]
Audience and channel: [who and where]
Approved assets: [product photos, screens, logo, script]
Shot 1 - hook: [first frame and reason to keep watching]
Shot 2 - action: [one visible change or demonstration]
Shot 3 - proof: [verified detail shown on screen]
End frame: [approved name, CTA space or clear conclusion]
Camera and sound: [only directions that help the story]
Keep unchanged: [product, character, brand and factual details]

If the workflow supports fewer shots, combine the action and proof. If it supports more, add only the beats that improve comprehension.

1. Write the outcome. State the video task and what the viewer should know or do afterward.
2. Map the shots. Give each shot one job and attach the references that make it accurate.
3. Review the draft. Paste the brief into your video workflow, inspect the preview and correct the shot that missed its purpose.

The shot plan is the prompt. Generation, editing and publishing remain separate steps.

Use the same structure for different tasks:

Product demonstration

Make a short video from the supplied travel mug photos. Hook: close view of the real lid. Action: a hand opens its actual latch and pours water. Proof: show the correct hardware and no invented feature. End: full mug profile with approved logo. Natural kitchen sound and steady camera.

App explainer

Use the approved booking screens. Hook: the user's goal in one caption. Action: show selecting a time and confirming it. Proof: the verified confirmation state. End: the approved product name. Keep screen order and on-screen wording exact; leave captions for human review.

Brand social clip

Use the real bakery exterior and pastry images. Hook: storefront sign in a vertical frame. Action: a hand places one loaf on the counter. Proof: a close view of the supplied pastry. End: the approved logo. Warm room ambience, no testimonial or invented offer.

Each example is a prompt plan. Actual video quality depends on the assets, model route and review process.

State the channel and preferred framing, such as vertical for a short social placement or horizontal for a product page, without assuming every model offers the same controls. Give a duration target only if the story needs it. Choose one coherent visual treatment; a long list of conflicting styles can obscure the action. Sound cues should name a source and moment, such as a latch click when the lid opens.

Rough idea: Show how our desk lamp works.

Completed prompt:

Use the approved lamp photos for a short product demonstration. Shot 1: close on the actual control dial. Shot 2: one hand turns the dial and the light changes. Shot 3: pull back to the full lamp on a bedside table. Keep the dial markings, color and logo unchanged. Use a quiet switch sound and finish with room for approved product copy.

Review the first draft for control accuracy, hand movement, sound timing and readable copy space. The example does not represent a generated output.

Supply only assets you have rights to use. Remove unnecessary personal information before uploading a brief or reference. Pexo's terms explain account conditions; inspect current product controls for any additional workflow-specific options.

Choose a matching task in Pexo's create directory or begin with its product-video workflow. Share the shot plan and authorized references, then review Pexo's direction and preview. Keep the parts that work and request a specific change to the missed shot. Approve facts, captions, sound rights and channel crop before publishing.