AI Tools / Video

Image to Video — First & Last Frame

Upload a first frame (required) and optional last frame—two still images, not a multi-image board. Describe motion between them and generate video with Seedance 2.0 or VEO 3.1.

Try Image to Video See pricing

First frame required · last frame optional (two slots only)
Seedance 2.0 & VEO 3.1 — not the same as Multi-images to Video
Motion prompts up to 4,000 characters; ~500 enough for most shots

Two frame slots: start (required) + end (optional)

JPEG, PNG, or WebP frame uploads

Seedance: adaptive ratio + optional last-frame mode

VEO 3.1: 16:9 or 9:16 · 4 / 6 / 8 seconds

Models

Which models support Image to Video?

Dashboard → Image to Video. Both models animate from your uploaded first frame (and optional last frame)—they do not accept arbitrary multi-image reference boards here.

Seedance 2.0 — duration 4–15 seconds (default 5); aspect uses adaptive ratio from your frames; quality 480p, 720p, or 1080p; optional sound; Advanced → With Last Frame / Not With Last Frame toggle when a last frame is uploaded
VEO 3.1 — duration 4, 6, or 8 seconds only (default 8); pick aspect 16:9 or 9:16 explicitly; quality 720p, 1080p, or 4K; optional sound; last-frame advanced toggle hidden—plan composition in the first frame
This is first/last frame mode (image_with_roles: frist frame / last frame)—for 1–9 loose references without frame roles, use Multi-images to Video instead

Inputs

What do you submit?

You provide up to two still images with defined roles—not a gallery of multiple reference images.

First frame (required): one JPEG, PNG, or WebP still marked Required in the UI. Generation is blocked with “First frame image is required” until it uploads.
Last frame (optional): one still in the second slot. Omit it and describe motion in the prompt instead (e.g. “slow dolly in”).
Prompt: up to 4,000 characters for camera path, subject action, lighting, and mood (~500 characters recommended).
Duration & quality: same ranges as Text to Video per model (Seedance 4–15s; VEO 4/6/8s).
Aspect ratio: VEO 3.1 shows 16:9 / 9:16 selector; Seedance uses adaptive ratio derived from your frames.
Sound (Advanced): With Sound / Not With Sound.
Optional: public gallery checkbox.

Frame uploads

What makes good first and last frames?

Because the model interpolates between defined start/end stills, frame quality and alignment matter more than prompt length.

Strong first frame

Use a sharp, well-lit still with clear subject separation. The first frame is always sent as first frame (start)—it anchors identity, pose, and composition.

Optional last frame for known endings

Upload a last frame (last frame) when you know the exact end composition—product fully revealed, door open, logo centered. Skip it when you only need motion from a single still.

Align perspective and lighting

Match scale, camera angle, and color temperature between frames to reduce morph artifacts. Keep the subject in similar positions when using both frames.

Prompting

Motion prompts for frame-based video

First frame only (no last frame)

Describe camera and subject motion explicitly: “Static product on table; slow orbit clockwise; soft studio light; shallow depth of field.” One clear camera verb works best.

First + last frame pair

Keep the prompt about transition: “Smooth morph from closed box to open box with product visible; maintain label readability; gentle ease-in-out.” Let the frames define start/end.

Camera path language

Use plain verbs: dolly in/out, pan left/right, orbit, crane up, handheld drift. Example: “Slow dolly-in on portrait; background bokeh; golden hour rim light.”

When to use this tool

Image to Video vs other tools

Image to Video is “defined start/end stills → motion clip.” It is not Text to Video (words only) and not Multi-images to Video (many references without frame roles).

Choose Image to Video when: you have a hero still (and optionally an end still) and need controlled motion between compositions—product reveals, morphs, portrait animation.
Choose Text to Video when: you have no frames yet and can describe the entire shot in words.
Choose Multi-images to Video when: you have several style/character/product guides (1–9 images) but no single defined last frame—Seedance only, different tool.
Choose Video to Video when: reference motion comes from uploaded video clips, not still frames.
Pipeline: Text to Image or Image to Image for stills → Image to Video for motion → export for ads or social.

Prompt blueprint

Frame-based video prompt template

Anchor: what the first frame already shows (do not contradict it)
Motion: one primary camera move or subject action
End state: if no last frame, describe where motion should finish
Lighting: keep consistent with both frames when using a pair
Duration fit: simpler motion for 4–6s VEO clips; richer arcs for longer Seedance runs
Not for: naming “image 1 / image 2 / image 3” roles—that belongs in Multi-images to Video

Examples

Example Image to Video outputs

Replace with your best first/last frame workflows to build trust from search.

Single-frame dolly

Only first frame uploaded; motion from prompt.

Box open reveal

First and last frame define closed → open states.

Portrait subtle move

Anchored face with gentle camera drift.

Product orbit

Studio product still with orbital motion.

FAQ

Is Image to Video the same as Multi-images to Video?

No. Image to Video uses exactly two frame roles—a required first frame and optional last frame. Multi-images to Video accepts 1–9 reference images without first/last roles and only supports Seedance 2.0.

Do I have to upload a last frame?

No. Only the first frame is required. Without a last frame, describe motion in your prompt (camera move, subject action).

What file types can I upload?

JPEG, PNG, or WebP still images in the frame upload slots.

Which models are available?

Seedance 2.0 and VEO 3.1 from the model picker. Settings differ: Seedance uses adaptive aspect ratio; VEO requires 16:9 or 9:16 and fixed 4/6/8s durations.

What happens if I forget the first frame?

The dashboard shows “First frame image is required” and blocks Generate.

When should I use Text to Video instead?

When you have no still frames yet and can describe the full shot in text only.

How long can my prompt be?

Up to 4,000 characters; the UI recommends around 500 for most shots.

Can I use outputs commercially?

Commercial usage depends on your plan and terms. Review pricing and policies before shipping at scale.

Animate from first & last frames

Upload your start still, optionally your end still, describe motion, and generate.

Try Image to Video View pricing