RedVideo Core
House image-motion route with managed audio.
- Page workflow
- First & Last Frame
- Duration
- 5s
- Output
- 720p
- Audio
- Model managed
FIRST + LAST FRAME / CONTROL BOTH ENDS
Guide AI video with a first and last frame
Set the visual state at the beginning and end of a shot, then describe the movement that should connect them. Supported RedVideo models treat both images as guidance while the selected workflow determines duration, format, access, and cost.
Content verification
Method: Model and workflow specifications are drawn from the current RedVideo configuration catalog. Live availability, supported settings, and generation quotes can change; confirm them in Studio before generating.
MODELS AND CAPABILITIES
Specifications below come from the workflows currently configured in RedVideo. Live availability and generation quotes remain visible in the studio.
Wan 2.7
Newest Wan route for text or a single starting frame.
- Page workflow
- First & Last Frame
- Duration
- 5s · 10s · 15s
- Output
- 720p · 1080p
- Audio
- Model managed
Veo 3.1 Fast
Fast Veo generation with a fixed 8-second output and text, frame, or reference-image guidance.
- Page workflow
- First & Last Frame
- Duration
- 8s
- Output
- Model managed
- Audio
- Model managed
Veo 3.1 Quality
Higher-quality Veo generation with a fixed 8-second output from text or first and last frames.
- Page workflow
- First & Last Frame
- Duration
- 8s
- Output
- Model managed
- Audio
- Model managed
Seedance 1.5 Pro
Seedance Pro with camera lock and optional audio.
- Page workflow
- First & Last Frame
- Duration
- 5s · 10s
- Output
- 480p · 720p
- Audio
- Optional where supported
Seedance 2.0 Standard
Full-resolution Seedance 2.0, including native 4K.
- Page workflow
- First & Last Frame
- Duration
- 4s · 5s · 6s · 7s · 8s · 9s · 10s · 11s · 12s · 13s · 14s · 15s
- Output
- 480p · 720p · 1080p · 4k
- Audio
- Optional where supported
Seedance 2.0 Fast
Faster Seedance 2.0 at 480p or 720p.
- Page workflow
- First & Last Frame
- Duration
- 4s · 5s · 6s · 7s · 8s · 9s · 10s · 11s · 12s · 13s · 14s · 15s
- Output
- 480p · 720p
- Audio
- Optional where supported
Seedance 2.0 Mini
Lower-cost Seedance 2.0 at 480p or 720p.
- Page workflow
- First & Last Frame
- Duration
- 4s · 5s · 6s · 7s · 8s · 9s · 10s · 11s · 12s · 13s · 14s · 15s
- Output
- 480p · 720p
- Audio
- Optional where supported
Define where the shot begins and where it should land
First-and-last-frame generation requires an opening image and an ending image. The pair gives the model two visual states to connect instead of asking it to invent both ends from a prompt alone.
Use compatible images that you own or are authorized to use. The studio validates each frame against the selected model’s file, dimension, and aspect-ratio requirements before submission.
- Choose a clear opening composition
- Prepare an ending frame with a plausible visual relationship
- Swap the complete frame assets when the direction should run in reverse
Prompt the change that happens between the frames
Describe the subject movement, camera path, environmental change, and pacing that should carry the first frame toward the last. A transition-focused prompt is usually more useful than repeating details already visible in both images.
The frames are guidance rather than a frame-perfect guarantee. Review continuity, object shape, camera movement, and the approach to the final composition, then revise one part of the direction at a time.
Choose a model for the duration and controls the shot needs
RedVideo currently exposes first-and-last-frame workflows on supported Core, Wan, Veo, and Seedance routes. Their duration, resolution, ratio, audio, and access options are not interchangeable.
Core currently uses a fixed five-second output for this workflow, while other supported models expose their own settings. The live studio is the source of truth for the active model, quote, and account eligibility.
Reduce unnecessary differences before generation
Frames with compatible framing, subject scale, lighting, and scene geometry give the model a more coherent transition problem. Large unrelated changes can still be attempted, but they leave more of the path between images for the model to infer.
If the goal is to preserve an identity, logo, or product shape, inspect the generated result carefully. AI generation is probabilistic and should not be treated as a guarantee of exact visual fidelity.
FREQUENTLY ASKED QUESTIONS
- What is first-and-last-frame AI video?
- It is a video workflow that uses one required image for the opening and another required image for the ending. A prompt directs the motion or transition between those visual anchors.
- Do I need to upload both frames?
- Yes. The workflow requires both a first frame and a last frame. Standard image-to-video is the better choice when you have only one starting image.
- Will the generated video match both images exactly?
- Not necessarily. Both images guide the output, but the generated motion and intermediate frames are probabilistic. Review important subjects, text, logos, and the final transition before using the result.
- Which RedVideo models support first and last frames?
- The current catalog includes supported Core, Wan 2.7, Veo 3.1, and Seedance routes. Available versions and settings can change, so check the workflow list shown for the selected model in the studio.
- Is first-and-last-frame generation free or unlimited?
- Access and credit requirements depend on the model, account, region, and current configuration. Eligible accounts may see separate Unlimited Core conditions, but that entitlement is not included for every account and optional audio can still use credits.
OPEN REDVIDEO
BUILD THE FIRST GENERATION
Select the model and workflow that fit the idea, review the live settings and quote, then start creating.