AI video shot list template for short-form edits

Plan the edit before the generations

A short-form edit needs contrast: wide and tight, still and moving, reveal and confirmation. Generating eight similar hero angles creates expensive repetition rather than a sequence.

Fields for every shot

A useful shot-list row contains enough information to reproduce the shot without turning into a screenplay.

• Function: hook, reveal, detail, transition, proof, or closing frame.
• Duration: the expected time in the final edit.
• Framing: wide, medium, close, macro, overhead, profile, or symmetry.
• Reference: the exact approved source image.
• Motion: one camera move and one dominant subject action.
• Continuity lock: the details that must not change.
• Sound cue: impact, rise, cut, voice line, or ambient transition.

Use inserts to protect the sequence

Macro details, reflections, packaging, hands, wheels, fabric, and environmental cutaways can hide difficult transitions and make an edit feel more expensive. They also require less complex motion than a full-body or full-product transformation shot.

Use one reusable row for every planned clip

A row should let another editor understand why a clip exists without reading the prompt. Record a shot ID, edit function, estimated duration, framing, approved source still, primary movement intention, continuity note, sound cue, and status. Keep the prompt and output link beside the row once production begins.

Run a selection gate before generation. If two adjacent rows have the same function and framing, revise one. If a row has no reference, continuity note, or place in the sequence, it is not ready for the next production stage.

• Shot ID and edit function: hook, reveal, detail, proof, transition, or close.
• Estimated duration and framing: wide, medium, close, macro, profile, overhead, or symmetry.
• Exact source still and the details that should remain recognizable.
• One primary movement intention plus a sound or transition cue.
• Status: planned, source approved, generated, selected, revised, or exported.

Map angles to purpose before adapting the vertical edit

In the Ford GT study, the master defined the vehicle and studio logic. The front hero offered presence and reveal, the rear offered a symmetry opportunity, the side profile clarified silhouette, the high angle changed visual scale, and tight carbon or branded details created insert candidates. These labels describe editorial intent rather than guaranteed model behavior.

For a vertical deliverable, review which part of each source still survives the narrower crop and whether the sequence still alternates scale and purpose. Reframe or replace a row when the subject, badge, wheel, or light structure becomes unreadable; do not keep a clip only because it was already generated.

Build UGC B-roll around evidence, not decoration

For a creator-style ad, separate the spoken performance from the coverage that lets an editor verify, pace, or clarify it. The A-roll row carries a specific message beat. Each B-roll row should then show a product fact, a plausible action, the usage context, or a clean transition rather than repeating a generic lifestyle close-up.

A cutaway does not turn a claim into proof. Use approved product references and claim language, keep labels and physical behavior accurate, and mark synthetic demonstrations or identities when disclosure is required. Reject a polished insert when the product, hand interaction, or implied experience is unreliable.

• Hook support: a readable product or situation frame that can open before the first spoken line.
• Product proof: label, scale, material, interface, or feature detail supported by an approved reference.
• Action coverage: one physically plausible use step with clear hand and product contact.
• Context: the environment or routine needed to understand when the product appears.
• Objection support: an approved detail that answers a real concern without inventing a result.
• Close: a stable product or creator frame with room for the approved CTA and disclosure.

Plan interchangeable coverage before generating variations

Give each important script beat one required row and, only where useful, one alternate with a different framing or action. Record which beat the alternate can replace. This creates edit options without turning the list into unrelated footage or multiplying every shot by default.

For every row, plan a clean entry, a readable hold, and a clean exit. Review the selected range with sound and captions, then confirm that the cutaway still matches the spoken line, product state, creator identity, and disclosure context. A shorter truthful range is more useful than a longer clip with drift.

Diagnostic table

Find the failed layer before regenerating.