# Burning Bridges Prompt Guide: Separate the Character, Room and Gesture Canonical: https://aiburningbridges.online/burning-bridges-prompt-guide/ Updated: 2026-10-11 Prepare the character and setting first. Use a motion reference for the performance, then change one failed element at a time. A Burning Bridges AI prompt is useful when it describes visible things: the subject, room, starting pose, hand action and camera relationship. A broad request to do the trend leaves the model to invent those details. Separate the character still from the motion instruction so revisions remain understandable. The prompts below are original creative suggestions for compatible advanced tools. They are not the source creator's prompts and have not been tested by a paid generation. This studio uses preset motion templates and has no custom prompt box. ## Start with one subject in a readable party room Build a still that already contains the outfit, light and composition you want to animate. Make the subject large enough to recognize, but leave room for the shoulders, hands and cup. A crowded room can supply atmosphere without competing foreground faces. Check the face and anatomy before motion. If a still has extra fingers or an uncertain lead, remove that problem at the image stage. A motion prompt cannot reliably repair every flaw in the input. ```text One consenting adult is the clear lead at a warm chandelier-lit indoor party. Head and torso framing, recognizable face and hairstyle, relaxed starting pose, visible hands, a plain cup at waist height. Background guests are less prominent than the lead. No captions or logos. ``` ## Give the toast one direction and a pause A controlled raise beside the face is easier to inspect than a cup waved across the lens. Describe the starting position and end pause. If the actual reference contains several turns, a short prompt will not remove them from the transferred performance. ```text Follow the reference toast movement. The subject slowly raises the cup beside their face, keeps the grip visible, pauses, then settles. Preserve the reference identity and outfit. Keep the cup in contact with the same hand. Do not add a spin, new camera cut, captions or logos. ``` ## Keep the searching gesture separate from lip sync The hand-near-the-brow action and a lyric-timed mouth movement are different requirements. This studio implements a motion-reference payload, not a verified lip-sync stage. Do not present a searching prompt as proof that a specific vocal line will match. ```text Follow the motion reference: the subject raises one hand near the brow, makes a small searching head turn and pauses. Keep part of the face visible. Preserve the character identity and clothing. No extra dance, dramatic zoom, new subject or text. ``` ## Prepare the animal body instead of hiding the mismatch For a dog or cat edit, first make a deliberate upright character image. Name the recognizable coat, ears and muzzle features, then specify visible front paws and a simple starting pose. The body needs to suit the performance before a human motion is transferred. Review the animal still against the original photo. A generic animal with party clothing may be attractive but not identifiable. Direct raw-pet generation remains unverified, so use the dog and cat preparation guides for the intended input. ## Match the revision to the failed part If the room looks right but the face changes, keep the setting and improve the subject reference. If the character is stable but hands break, simplify the gesture. If the model animates a background guest, make the lead more prominent in the still. Keep a record of the exact image, reference, quality and one change made for each take. A new generation can vary with identical inputs. Those records help compare attempts, but they do not turn an untested prompt into a reliable guarantee. 1. Face drift: use clearer identity detail and less occlusion. 2. Cup deformation: simplify the grip and hand path. 3. Wrong lead: reduce competing foreground faces. 4. Lost framing: prepare more space around the moving arm. 5. Unwanted sound: check the reference and downloaded audio track. ## Finish the edit after the performance reads clearly Inspect the clip without music first. Check the face, then the hands, then the scene continuity. Higher quality can reveal detail but cannot solve contradictory instructions or an unsuitable reference. Add permitted audio only after the visual sequence works. A Burning Bridges prompt should give you something specific to inspect. Use one clear subject, a prepared room and a reference with readable gestures. Keep claims about output quality and timing limited to the test you actually ran. ## Questions about this guide ### Can I paste these prompts into the studio? No. The studio uses preset templates. These examples are for a separate compatible advanced workflow. ### Are these the original Burning Bridges prompts? No. They are original preparation suggestions, and the source creator's settings have not been confirmed. ### Does writing 4K in a prompt produce 4K? No. Resolution is a provider setting and a delivered-file property. Select 720p, 1080p or 4K in the quality menu; writing a resolution into the prompt does not change the requested output. ## Before your next take Write a Burning Bridges prompt around a visible subject and action. Prepare the image, transfer a checked performance, and revise the failed element rather than the entire scene. [Explore the studio](https://aiburningbridges.online/burning-bridges-prompt-guide/#create)