Identify what should be retained.
Write down whether the priority is the lead, the motion or the camera. Use references you have permission to transform.
REFERENCE VIDEO / CHARACTER / NEW SCENE
Genjutsu AI: Plan a Reference-Based Video
Decide what your reference should guide: the person, the movement or the setting. Compare the editing idea with a fresh scene and prepare a clear reference brief.
Authorized photo + optional videoEditable scene promptYour existing credits
Preview
Your next take will appear here.Orange-booth duet as a reference-workflow illustration · 720p · 10s · generated demo
Recent takes
View historyNo generated videos yet.
Your inputs are saved while you work.What does Genjutsu AI mean?
Higgsfield Genjutsu is a separate video-editing product with Motion Transfer and Object Swap workflows. Burning Bridges AI does not run that product. This page explains the distinction and provides an editable reference-based generation workflow.
Read the original Genjutsu product guideReference generation and source editing differ
Here, a supported model creates a new video from your instructions and references. It does not guarantee preservation of untouched source pixels, precise choreography or a local object replacement. Dedicated Video to Video remains marked as planned; supported video references are available in multi-reference generation.
Choose material that supports your goal
Use a short authorized video with a clearly visible action if timing matters. Use a separate portrait for the new lead and name it explicitly in the prompt. Reference-video duration affects the credit quote, so review the full estimate before submission.
REFERENCE VIDEO / CHARACTER / NEW SCENE
Prepare a Reference Brief Before Generating
Assign separate roles.
Use the supported model’s photo and video-reference inputs. Describe which input guides identity and which guides movement.
Compare the whole result.
Check the new scene against the source. Exact local edits and source-frame preservation are not promised.
THE CHARACTERS IN THIS TAKE
Give each reference one job.
This existing Hotel Lobby demonstration illustrates a scene you can describe. It is not output from the official Genjutsu product and does not demonstrate object swapping or exact motion transfer.
Give each reference one job.
A portrait can guide the lead’s appearance. A supported video reference can guide movement and timing. Explain which details matter and inspect the new scene against those references.
- Example file
- 720 × 1280
- Length
- 10 seconds
- Frame
- 9:16 portrait
- Sound
- Original generated audio
YOUR STARTING PROMPT
A Reference Video Prompt You Can Adapt
This editable starting prompt is already loaded in the workspace. It is a new direction you can adapt, rather than the exact generation settings of the displayed demo.
Generate an original scene using only authorized adult subjects. Reference video 1 guides the pacing, gestures and camera direction; reference image 1 defines the lead person, face and outfit. Describe a warm recording-booth performance with one shared microphone and a readable head nod and hand gesture. Keep the lead’s identity stable and the action simple. Interpret the reference as creative guidance for a new shot, not an exact edit of every source frame. No public-figure impersonation, existing soundtrack recreation, captions or logos.Edit the prompt in the workspace
Genjutsu AI Questions
Is this the official Genjutsu AI tool?
No. This is an independent guide and a separate Burning Bridges AI reference-generation workflow.
Can I use a video reference here?
Yes, in multi-reference generation with a supported Seedance model. Model-specific reference limits and credit quotes apply.
Can this replace an object while keeping every other source pixel?
That dedicated local-editing feature is not offered here. This workflow generates a new scene.
Is the displayed clip a Genjutsu result?
No. It is an existing fictional Hotel Lobby demo, labeled as an illustration.