First and Last Frame to Video
Guide a video with a starting image and an optional ending image. Choose a frame-compatible model in Kivue and describe the transition between the supplied compositions.
Credit cost depends on the selected model and settings. Check the amount shown before submitting.

FROM INPUT TO PROJECT
What can you create with First and Last Frame to Video?
Guide a video with a starting image and an optional ending image. Choose a frame-compatible model in Kivue and describe the transition between the supplied compositions.
VISUAL REFERENCES
Browse 3 ideas from the Kivue prompt library
Open a cover to inspect its template. These are library references, not a benchmark of this tool or model.



What output settings does First and Last Frame to Video support in Kivue?
13 model configurations are available for this entry. Each row describes one configuration; do not combine settings from different rows.
| Model option | Output size | Aspect ratio | Duration |
|---|---|---|---|
| Seedance 2.5 | 480p, 720p, 1080p | 1:1, 9:16, 16:9 | 4–30 seconds |
| Seedance 2.0 | 480p, 720p, 1080p | 1:1, 9:16, 16:9 | 4–15 seconds |
| Seedance 2.0 Mini | 480p, 720p | 1:1, 9:16, 16:9, 4:3, 3:4, 21:9 | 4–15 seconds |
| Kling 3.0 | 720p, 1080p | 1:1, 9:16, 16:9 | 5–15 seconds |
| Vidu Q3 Pro | 720p, 1080p | 1:1, 9:16, 16:9, 4:3, 3:4 | 1–16 seconds |
| Vidu Q3 Turbo | 540p, 720p, 1080p | 1:1, 9:16, 16:9, 4:3, 3:4 | 1–16 seconds |
| MiniMax H3 Max Turbo | 480p, 768p | 1:1, 9:16, 16:9 | 5–15 seconds |
| MiniMax H3 Max | 480p, 768p | 1:1, 9:16, 16:9 | 5–15 seconds |
| MiniMax H3 | 768p, 2k | 1:1, 9:16, 16:9 | 4–15 seconds |
| Wan 3.0 Prime | 480p, 720p, 1080p | 1:1, 9:16, 16:9, 4:3, 3:4 | 2–30 seconds |
| Wan 3.0 | 480p, 720p, 1080p | 1:1, 9:16, 16:9, 4:3, 3:4 | 2–30 seconds |
| Wan 2.7 | 720p, 1080p | 1:1, 9:16, 16:9 | 2–15 seconds |
| Kling 4.0 | 720p, 1080p | 16:9, 9:16, 1:1 | 3–30 seconds |
These are Kivue integration settings, not a claim about every feature offered by the model developer. Reference modes can impose additional limits.
Source: Kivue current model configuration. The generator quote remains the price to review before submission.
How should you prepare a First and Last Frame to Video request?
Example prompt to adapt to your input:
Start with the closed box in the first image and end with the open box in the second. Keep the camera fixed and the product in the same place.
The endpoint images constrain composition, not every intermediate frame. Avoid incompatible viewpoints, object counts or lighting between endpoints.
Start with 1 result. Keep the same reference and change one setting at a time when comparing compositions or movement.

How do Kivue and alternative services organize this workflow?
This 3-service comparison covers public entry points and workflow design. It does not rank output quality, speed or price.
| Service | Entry or inputs | Workflow described on the source page |
|---|---|---|
| Kivue | Prompt and model-supported references | Seedance 2.5 lists 480p, 720p, 1080p, with a configured duration of 4–30 seconds. Select a model and check the task quote before submitting. |
| Pollo AI | Text-to-video entry | Its public navigation links model pages and separate image and video tools. |
| Clipdance | Text, image and reference-video entries | Its public navigation separates video workflows, image workflows and AI tools. |
Public competitor pages checked October 30, 2026. A listed entry is not evidence of an identical underlying model. Kivue-specific limits are documented in the configuration above.
How do you use First and Last Frame to Video in Kivue?
Follow these 3 steps. Review the settings for this task rather than assuming a previous task used the same parameters.
- Prepare the input. Describe the scene, subject movement and camera direction. Add references only when supported by the selected mode.
- Choose the output. Check the available settings and displayed credit quote. If the workbench is unavailable, it cannot accept a new request.
- Inspect and save. Wait for the completed task, review the result against the input, and save the version you intend to use.
What should you check before publishing the result?
Inspect images at 100% zoom. For video, inspect the start, middle and end; a correct poster frame does not establish that the whole clip is correct.
Look for changes in product shape, face identity, text, object count and continuity during movement. A higher output resolution does not establish that these details are accurate.
Frequently asked questions about First and Last Frame to Video
These 4 answers cover input preparation, output checks, cost and availability.
Which resolutions and aspect ratios can I select for First and Last Frame to Video in Kivue?
Seedance 2.5 lists 480p, 720p, 1080p, with a configured duration of 4–30 seconds. Its listed ratios include 1:1, 9:16, 16:9. These settings describe this Kivue configuration. Other variants and input modes can differ; check the form for the allowed combination.
How should I write a prompt for First and Last Frame to Video for a product campaign?
The endpoint images constrain composition, not every intermediate frame. Avoid incompatible viewpoints, object counts or lighting between endpoints.
How many credits will one First and Last Frame to Video generation use?
The form calculates a quote from the selected model, output settings and quantity. Review that amount before submitting. A resolution listed on this page is not a fixed-price promise.
Does First and Last Frame to Video reproduce my product reference without changing any details?
No exact-match guarantee applies. Reference material guides generation, but shape, print, text or identity can change. Compare the output with the source before commercial use.