Generate video with synchronized audio from text, ordered images, or an existing clip. This guide covers the controls that matter in Sequencer.
Generate with FLUX.3
Explore the modes
Sequencer
×
Black Forest Labs
·
Updated August 2026
·
12 min read
FLUX 3 is Black Forest Labs’ first multimodal video model. It generates motion and scene-aware audio together, so footsteps can land with the frame, dialogue can match the performance, and ambience can change with the space.
This guide focuses on the choices that shape a real shot: which input mode to use, how ordered images become keyframes, how to continue a clip without exposing the seam, and how to spend less while you iterate.
Video and audio in one generation
Sound is part of the shot
Not a second pass or a separate model
Direct dialogue, foley, room tone, effects, music, or deliberate silence in the same prompt as the visual action. FLUX 3 can use text alone, ordered images, or an existing clip as the starting point.
Preview status
FLUX 3 is currently an early-access preview. Video generation, ordered image keyframes, drafts, and continuation are available; dedicated video editing and mixed image-plus-video Omni Reference are still listed as coming soon.
FLUX 3 specifications
These are the practical limits for the current Sequencer integration.
Duration
5–20 seconds
Choose any whole-second duration in that range.
Aspect ratio
7 formats
From cinematic 21:9 through vertical 9:16.
Native audio
On by default
Dialogue, effects, ambience, and music direction.
Keyframes
Up to 10
Opening frame, end frame, or an ordered storyboard.
Choose a generation mode
Choose the mode based on what you already have: an idea, visual anchors, or a clip that needs to keep going.
Text to video
01
FLUX.3 / FLUX.3 Draft
Start from a shot direction. Best for ideation, establishing shots, inserts, and sound-aware moments without existing artwork.
Image to video
02
FLUX.3 / FLUX.3 Draft
Attach one opening image, a start/end pair, or up to ten ordered images that form a storyboard through the shot.
Continuation
03
FLUX.3 Continuation
Attach an MP4 and generate what happens next while carrying subject, framing, motion, and sound logic forward.
Text-to-video starter prompt
Wide establishing shot of a rain-soaked night market beneath warm paper lanterns.
Two friends hurry through the alley as the camera tracks backward at walking pace.
Steam drifts from a food cart and raindrops form ripples in the foreground puddles.
Audio: close footsteps through water, rain on canvas awnings, quiet crowd ambience,
a vendor calling in the distance. Natural sound only, no music.
Style: grounded 35mm cinema, warm practical light, deep blue evening shadows.
Generate a FLUX 3 clip
Use Draft while the direction is still moving, then switch to a full render once the shot works.
1
Open a project and choose Video
Create or open a Sequencer project, open the generation panel, and switch the media type to Video.
2
Choose a FLUX.3 model
Pick FLUX.3 Draft to explore a new text or image-driven shot. Choose FLUX.3 for a final-quality clip, or a Continuation option when starting from video.
3
Add visual inputs if you have them
One image becomes the opening frame. Two images define the opening and closing frames. Three to ten images become a storyboard in attachment order.
4
Set the format
Choose a 5–20 second duration and your delivery aspect ratio. Full renders support 720p and 1080p; Draft renders at 720p.
5
Direct picture and soundtrack
Describe subject, action, camera, environment, timing, and audio. Keep the shot coherent enough to happen inside the duration.
6
Review it as one take
Listen with sound on. Judge performance, motion, and audio together, then add the keeper to your sequence or refine the prompt.
Use ordered images as keyframes
These inputs become actual frames of the clip—not just loose style references.
1 image
Pins the opening frame. Your prompt directs everything that follows.
2 images
Pins the opening and closing frames. FLUX 3 creates the transition.
3–10 images
Forms an ordered storyboard. Set an explicit duration so intermediate frames can be spaced across it.
Image-to-video motion prompt
The camera rises slowly above the alley while the two characters move deeper into frame.
Lanterns sway in the rain and reflections stretch across the wet stones.
Preserve both characters, their clothing, the environment, and the warm-to-cool color contrast.
Audio: rain, soft footsteps, cloth movement, distant market ambience. No music.
Prompt for change
When the images already define appearance and composition, spend the prompt on motion, timing, camera behavior, and sound. Re-describing every visible detail can introduce conflict and drift.
Continue an existing clip
Continuation carries the final frames, momentum, and sound logic of your source into newly generated footage.
Original clip
FLUX 3 continuation
Attach one MP4 up to 15 seconds and 50 MB. Prompt only what happens after the final frame, while carrying screen direction, camera speed, subject momentum, and audio perspective across the join.
Continuation prompt
Continue without a cut. They leave the alley and enter a crowded covered market.
Keep their pace, screen direction, wardrobe, camera height, and handheld movement consistent.
The rain becomes quieter beneath the roof as crowd chatter and cooking sounds grow louder.
Write a useful FLUX 3 prompt
A useful prompt tells the model what changes over time—not only what the first frame looks like.
Name the dialogue, room tone, surface sounds, distance, and whether music should be absent. “Quiet” is still a sound direction—and often a more useful one than generic cinematic music.
Use Draft while you iterate
Drafts render faster at 720p and cost about a third of a full HD render.
1. Draft
Test composition, timing, and sound direction at 720p.
2. Review
Judge the picture and soundtrack as one take.
3. Full render
Run the locked direction at the resolution your edit needs.
Cost per output second
Mode
Full render
Draft
Text / image to video
Full render
$0.17/s HD · $0.29/s FHD
Draft
$0.06/s HD
Video continuation
Full render
$0.43/s HD · $0.54/s FHD
Draft
$0.12/s HD
Black Forest Labs API list pricing as of August 6, 2026. Sequencer credits and partner pricing can differ; use the cost estimate shown before generation.
Generate with FLUX 3
Start with a 5-second draft, direct both picture and sound, and add the take that works to your sequence.