Veo 3.1 Fast workflow

Reference to Video AI Generator

Upload a small set of visual references, describe the motion you want, and generate one coherent clip without rebuilding the subject and style in every prompt.

1-3 reference images
16:9 output
Veo 3.1 Fast
300 credits per run

Start with references

From references to one moving scene

Choose consistent references

Upload clear images that show the subject, environment, wardrobe, or visual style you want the model to preserve.

Describe motion, not the stills

Tell the model what changes over time: subject action, camera movement, timing, and the transition between views.

Review and generate

Confirm the reference set and prompt, then run the shared Veo workflow and download the completed clip.

Create your reference-guided video

Reference mode is selected for you. Add 1-3 images, write the intended motion, and keep the required 16:9 Fast settings.

How to get a more consistent result

Keep one clear subject. Use references with the same person, product, or character and avoid conflicting identities.
Keep lighting and style compatible. References with similar color, lens, and lighting cues are easier to combine into one shot.
Write one camera plan. Specify a single movement such as a slow orbit, push-in, tracking shot, or controlled transition.
Remove unnecessary references. More images are not always better; keep only the frames that provide distinct, useful guidance.

Current workflow limits

Reference-to-video uses Veo 3.1 Fast, accepts 1-3 PNG, JPG, JPEG, or WebP files up to 10 MB each, and requires a 16:9 output. Invalid or missing inputs fail before credits are spent.

Uploaded references are used to complete the requested generation. Do not upload confidential material or images you do not have permission to use.

Reference-to-video questions

How is reference-to-video different from image-to-video?

Image-to-video animates one source image or a start and end frame. Reference-to-video uses 1-3 images as broader guidance for subject identity, composition, and style.

What can I upload?

Upload 1-3 PNG, JPG, JPEG, or WebP images, no larger than 10 MB each. Clear, compatible references usually produce the most stable result.

What should the prompt describe?

Focus on motion and time: what the subject does, how the camera moves, the pace, and how the scene should transition.

Which model and aspect ratio are used?

The current reference workflow uses Veo 3.1 Fast and locks the output to 16:9 for compatibility.

From references to one moving scene

Choose consistent references

Upload clear images that show the subject, environment, wardrobe, or visual style you want the model to preserve.

Describe motion, not the stills

Tell the model what changes over time: subject action, camera movement, timing, and the transition between views.

Review and generate

Confirm the reference set and prompt, then run the shared Veo workflow and download the completed clip.

How to get a more consistent result

Keep one clear subject. Use references with the same person, product, or character and avoid conflicting identities.

Keep lighting and style compatible. References with similar color, lens, and lighting cues are easier to combine into one shot.

Write one camera plan. Specify a single movement such as a slow orbit, push-in, tracking shot, or controlled transition.

Remove unnecessary references. More images are not always better; keep only the frames that provide distinct, useful guidance.

Current workflow limits

Reference-to-video uses Veo 3.1 Fast, accepts 1-3 PNG, JPG, JPEG, or WebP files up to 10 MB each, and requires a 16:9 output. Invalid or missing inputs fail before credits are spent.

Uploaded references are used to complete the requested generation. Do not upload confidential material or images you do not have permission to use.

Reference-to-video questions

How is reference-to-video different from image-to-video?

Image-to-video animates one source image or a start and end frame. Reference-to-video uses 1-3 images as broader guidance for subject identity, composition, and style.

What can I upload?

Upload 1-3 PNG, JPG, JPEG, or WebP images, no larger than 10 MB each. Clear, compatible references usually produce the most stable result.

What should the prompt describe?

Focus on motion and time: what the subject does, how the camera moves, the pace, and how the scene should transition.

Which model and aspect ratio are used?

The current reference workflow uses Veo 3.1 Fast and locks the output to 16:9 for compatibility.