Burning Bridges AI Video Prompt: One Photo, One Clean Party Shot

A useful Burning Bridges AI video prompt should establish six things in order: the referenced person, the original setting, the camera, three timed actions, one stable prop, and the sound inside the room. Put the identity anchor first. Save broad mood words for after the physical shot is clear.

Here is a complete starting prompt for a ten-second office-afterparty scene:

Use the uploaded image as the reference for the adult lead subject. Keep
their face, hairstyle, skin tone and outfit recognizable throughout.

Continue into an original copper-lit office lounge after closing time.
Use one continuous 16:9 medium shot. The handheld camera moves backward
slowly and keeps the subject near the center. The elevator closes behind
the subject; they take three relaxed steps, give one small wave toward a
coworker near the wall, then straighten the front of their jacket. The
camera settles as they pause beneath a copper wall light. Keep the hands,
clothing and background guests stable. No cuts or fast choreography.

Audio: an original low-key rhythm, one elevator chime, footsteps and
distant conversation. No commercial song, recognizable lyrics, artist
voice, celebrity likeness, brand marks, captions or on-screen text.

Load the same idea with the Office Afterparty preset. The Burning Bridges AI Video Generator makes a new scene from your subject reference and prompt, with original AI-generated audio. It is not an official music video, does not provide the original song or Drake's voice, and does not attempt a frame-by-frame recreation. AI Live Generator is not affiliated with or endorsed by Drake.

Why this prompt is ordered this way

The first sentence tells the model which details must carry from the reference into motion. “Keep the person the same” is vague; face, hairstyle, skin tone, and outfit name the details you will actually inspect.

The next paragraph describes a place without borrowing a named venue or set. “Original copper-lit office lounge” is enough to establish warm party light while leaving the generator room to create its own architecture and guests.

Then the prompt chooses one camera move. The camera retreats while the subject advances, so the face stays in view. The action is written as a sequence rather than a pile of simultaneous verbs: elevator closes, three steps, wave, jacket adjustment, settle.

Finally, the sound paragraph requests sources you could hear in that location. It does not use an artist name as shorthand for tempo or delivery.

Write actions the camera can actually see

“Looks confident” is an interpretation. “Takes three relaxed steps, waves once, then straightens the jacket” is visible direction. Use verbs that produce a change on screen, and give each one a place in the sequence.

This compact pattern works for most prompts:

Reference: preserve [identity details] from the uploaded subject photo.
Setting: an original [specific place] with [two visual details].
Camera: one continuous [framing] with [one camera move].
Action: first [beat one], then [beat two], finally [beat three].
Continuity: keep [hand/prop/clothing/background relationship] stable.
Audio: original [rhythm/ambience/physical sounds].
Exclude: [commercial music, lyrics, imitation, logos and text].

Do not spend half the prompt listing adjectives such as cinematic, epic, premium, viral, and iconic. One material, one light source, and one camera behavior tell the model more.

Give each hand one job

Hands drift when a short prompt asks them to wave, point, hold a prop, change grip, and touch clothing in the same take. Assign one brief gesture to one hand, then let both hands settle before the closing pose.

If the hand is more important than the entrance, simplify further:

The subject is already standing still. Their right hand straightens the
front edge of the jacket once, then returns to their side. The left hand
stays relaxed and visible. The face remains fully visible.

That is easier to judge than “the subject gestures confidently.”

Keep the crowd in a supporting role

Background guests should provide depth and movement without becoming extra lead characters. Try this language:

Four or five distinct adult guests remain behind the subject, softly out
of focus. They make small independent movements and never cross in front
of the subject's face or interact with the subject.

Avoid asking for a packed room if your main goal is identity consistency. More bodies mean more faces, hands, and occlusions competing inside the same short shot.

Two prompt variations that stay original

For a fantasy version, change the setting and action—not just the color:

Continue into an original candlelit stone banquet hall with fictional
adult guests. The subject walks between two tables, notices an empty
chair, gives one amused look toward the camera, and sits. Use one slow
medium-wide handheld shot. Audio: an original soft hand-drum rhythm,
fireplace crackle, chair movement and quiet table conversation.

For a quieter ending, remove the party completely:

Continue onto a rain-cleared rooftop at sunrise. The subject faces the
skyline, places a blank folded note on a concrete ledge, pauses, then
walks toward the warm light. One slow lateral camera move. Audio: wind,
distant traffic, footsteps and an original two-note synth motif.

The Fantasy Banquet preset and Rooftop Goodbye preset load fuller versions of those briefs.

Fix the smallest possible part

Use 480P to test whether the prompt's structure works, then choose 768P when the shot is stable enough to benefit from more detail. Keep the same duration while comparing prompt changes. If you change the photo, action, camera, length, and resolution together, you will not know which edit solved the problem.