User Guide
Muse Video User Guide
Learn how to turn a prompt or an image into a short video with sound. See the inputs and results from real examples, then try the same workflow with your own scene.
Create a Muse Video →
On this page
Create your first video
Choose a starting point
Use Text to start with a prompt, Image to animate a still, or References to show the generator the look you want.
Describe the shot
Name the subject and action. Add the setting, camera movement, and the sounds you want to hear.
Generate and watch with sound on
Check the displayed credit cost before submitting. Watch the whole clip to see whether the motion, identity, and audio hold up.
Refine one detail
If the result misses something, change that detail in the prompt and run it again. Keep the rest of the prompt so you can compare the results.
Model overview
Key capabilities
Start from text, animate a still, or guide the scene with image references. Describe sound in the same prompt as the action.
Model specifications
- Inputs
- Text prompts and image uploads
- Aspect ratios
- 16:9 · 9:16 · 1:1 · 4:3 · 3:4 · 3:2 · 2:3
- Audio
- Describe dialogue, ambience, or effects in the prompt.
Getting started
Choose a creation mode
- Text: start with a written scene.
- Image: upload one still to animate.
- References: upload images that guide the subject or look.
Write effective prompts
Start with one subject, one action, and one setting. Add a camera direction when it matters. For audio, name the sound and what makes it.
Features and usage
Image references
In References mode, upload images that show the subject, outfit, product, or setting you want. Use the prompt to describe how the inputs should appear in the result.
Upload and mention references
Upload your images first. Type @ in the prompt box to select an uploaded image, then say what to take from it. For example, use one image for the face and another for the outfit.
Character references
Use a clear portrait for the face and a full-body image for clothing and proportions. If the character changes clothes or location, repeat which features should stay the same.
Direct the action
In Muse Video mode, describe movement in the prompt. Name where the subject starts, what it does, and where it ends. The Muse Video reference uploader accepts images.
Camera movement
Name the shot size and one camera move. For example, ask for a close-up with a slow push-in or a wide shot that tracks alongside the subject.
Storyboards
Break a clip into beats. For each beat, name the framing, visible action, and any sound that needs to land with it.
Resolution
Check the actual output in Video Details. The Muse Video editor exposes aspect ratio, while output resolution is handled by the model. Export options are listed by plan on the pricing page.
Long-form generation
Build a longer scene from short clips. Repeat the character, lighting, and location cues across prompts, then join the clips in a video editor.
Revise a video
Change the prompt and generate a new version. To trim or combine finished clips, download them and use a video editor.
Recreate a clip
Open an example on the homepage and select Recreate to send its prompt to the generator. Upload any reference images you want to reuse.
Visual style
Describe the lighting, colors, materials, and lens effect you want. Use a reference image when the look is easier to show than describe.
First frame
Use Image mode to start from a still. Describe what should move after the opening image, such as a turn of the head or a slow camera pullback.
Arrange keyframes
Muse Video Image mode takes one starting image. To plan several visual beats, write them in order in the prompt or generate separate clips.
First and last frames
In Muse Video mode, upload the starting image and describe the ending in the prompt. The editor does not expose a separate last-frame input for this model.
Ending frame
Describe the last composition directly. For example, ask the camera to settle on the bottle label as the background falls out of focus.
Intermediate keyframes
Describe intermediate beats with short time ranges or an ordered shot list. Use separate clips when an exact intermediate frame matters.
Muse Video examples
Compare the prompt and reference images with the finished clip.
Direct the camera with text
Describe the subject, where it starts, and how the camera follows it. In this example, the car and the racetrack give the action a clear starting point.
Prompt
Generated result

Keep product details in view
Use a product reference when the shape, label, and finish matter. Tell the model which details should stay the same as the camera moves.
Reference materials
Prompt
Generated result

Keep a character consistent across shots
Pair a portrait with a turnaround to show the face and outfit from different angles. Use the prompt to describe what changes from shot to shot.
Reference materials
Prompt
Generated result

Break a short story into beats
Give each beat one action. This prompt moves from an entrance to a cake surprise, then to dancing and a final group moment.
Prompt
Generated result

Change the setting while keeping the same face
Use one character reference and specify what changes in each cut. Watch the output for changes in facial features as the location and outfit switch.
Reference materials
Prompt
Generated result

How to get closer to the result you want
- The subject changes mid-shot
- Use a clearer reference image and reduce the number of changes in the scene.
- The camera does the wrong thing
- Ask for one camera move. For example, use “slow push-in” in place of “dynamic cinematic movement.”
- The sound does not match the action
- Name the sound source and the moment you want to hear it. Keep spoken lines short.
This editor’s settings
With Muse Video selected, choose the aspect ratio in the editor. References mode accepts up to 4 images in the upload form. Image mode takes one starting image.
Check the cost shown on the Generate button before submitting. Compare credit packs, export resolution, and download options on the pricing page.
Questions and troubleshooting
Which workflow should I start with?
Choose Text when you have a scene in mind, Image when you want to animate one still, or References when several images define the subject or look. Start with the input that matters most to the shot.
Do I need to upload an image?
You can generate from a text prompt alone. Add an image when a character’s face, a product’s shape, or the opening composition needs to match an existing image.
How should I count reference inputs?
The Muse Video upload form accepts up to 4 images in References mode. Image mode takes one starting image. Give each reference a clear job, such as defining the face, outfit, or setting.
How long should my prompt be?
Start with one or two sentences for a single shot. Include the subject, action, and setting. Add camera and sound details where they matter. For several shots, put the beats in order so the timing is easy to follow.
Can I upload a music track or a sound-effects file?
Muse Video’s reference uploader accepts images. Describe dialogue, music, and sound effects in the prompt. If you need a specific track, add it to the finished clip in a video editor.
Do I need to turn on separate sound or storyboard switches?
Write sound cues and shot order directly in the prompt. To use a storyboard image, add it in References mode and explain which panels describe the opening, middle, and ending.
What should I change when the result misses the brief?
Pick one issue to fix first. If the camera moves too fast, specify a slower move. If the subject changes, simplify the action and use a clearer reference image. Keep the rest of the prompt for comparison.
Should I keep adding more detail to a prompt?
Add details that change what the viewer sees or hears. Replace vague phrases with concrete directions, such as a close-up of the label or footsteps on wet gravel. Remove directions that conflict with each other.
Why does the editor reject my file?
Check the selected model, file type, and upload count. For Muse Video, use a supported image format such as JPG, PNG, or WebP and keep each image within 10 MB. Image mode accepts one file.
Where can I review a completed generation?
Sign in and open History in the editor’s preview area. Select Muse Video if you were working with another model, then open the finished clip to review it.
Try your first scene
Write a short prompt or start with an image you want to animate.
Generate a Video



