Uploaded start frame
This still was uploaded first to guide the opening.
FIRST + LAST FRAME VIDEO
First and Last Frame Video Generator
Upload the image that opens the clip, optionally add the image that closes it, and describe how the scene should move between them.
Reference Image
Upload start frame
Upload end frame
The start frame opens the video, the end frame closes it, and the video transitions between the two.
AI enhancement
0 / 2,000
Two abstract stills, one generated transition
We tested these two in-house ribbon images on October 2, 2026. Veo 3.1 delivered an 8-second clip after the content checks. The motion and extra glow show how the model interprets the space between the endpoints.
Settings used
Image to video · Veo 3.1 · 8 seconds · 720p · 16:9.
Original prompt
Let the abstract violet light ribbons flow slowly from the uploaded start frame to the uploaded end frame. Keep the camera locked, the background near black, and the glow soft. Use smooth continuous motion with no text, no cuts, and no new objects.
A start frame, an optional end frame, and a motion brief
The second upload is optional. A single start image still works as ordinary image-to-video; when both images are present, the model generates a transition between them.
1
Upload the start frame
Choose a still you own or have permission to use. This image opens the clip and gives the model the starting composition.
2
Add an optional end frame
Choose a second still for the closing moment. This option appears only for models that support an end frame.
3
Describe the movement
Guide the camera, subject action, pacing, and lighting between the two images. Keep the prompt about the transition rather than restating every pixel.
4
Generate and review
The model fills in the intermediate frames. Review the result, then download it from your history if it fits the brief.
Two images guide the path; the model creates the middle
First-and-last-frame generation gives the render a beginning and an ending. It does not force every intermediate frame or guarantee exact identity, geometry, or timing.
The first image opens the clip
The uploaded start frame is the opening reference. A clear subject, stable framing, and enough room for motion usually give the model a stronger starting point.
The second image closes the clip
When you add an end frame, it is treated as the destination of the transition. It is optional, and the control is available only for models that support end frames.
Prompts guide motion, not exact pixels
Use the prompt to describe movement, camera direction, and continuity. The generated middle is an interpretation; matching the two uploaded images perfectly is not guaranteed.
Use images you can legally use
Upload your own work or an image you are licensed to animate. Content checks still apply before generation.
Settings follow the selected model
Available durations, resolutions, aspect ratios, and credit costs vary by model. The workspace only offers supported combinations.
PROMPT IDEAS
Two safe starting points to adapt
These are prompt ideas, not verified outputs. Try them with your own start and end frames, then adjust one motion detail at a time.
- 01
A restrained transition
Create a smooth, restrained camera move between the uploaded start and end frames. Preserve a coherent layout, subject placement, and lighting as the scene transitions naturally.
Untested ideaMotion direction - 02
A soft lighting transition
Move slowly from the uploaded start frame to the uploaded end frame with a gentle camera push and a soft change from morning light to warm afternoon light. Keep the subject grounded and the transition calm.
Untested ideaLighting continuity
First-and-last-frame video questions
Do I need two images?
No. The end frame is optional. Upload only a start frame for a regular image-to-video clip, or add a second image when you want to guide the closing moment.
What does the second image do?
It gives the model a target for the end of the clip. The start frame opens the video, the end frame closes it, and the model generates the transition between them.
Will the middle match my images exactly?
No result can promise that. The model creates the intermediate frames, so subject identity, geometry, and timing can drift. A prompt can guide the motion and continuity but cannot force every intermediate frame.
Which models can use an end frame?
The page shows image-to-video models that support an end frame. Models without that support are hidden for this workflow; the available duration, resolution, and aspect-ratio controls still depend on the model you choose.
How many credits does a clip cost?
The exact credit cost depends on the selected model, duration, resolution, and other supported settings. The workspace shows the current estimate before you submit. If a content check blocks a request before generation starts, no credits are charged; later technical failures follow the current refund policy.
Can I use any image?
Upload only images you own or are licensed to use. Do not use private or identifiable material without permission, and remember that every upload and generated clip remains subject to the platform and model provider terms.