Shape a Seedance 2.5 video draft from a written scene or one starting image, then choose duration, resolution, frame shape, and audio before you continue.
Seedance 2.5 is a flexible choice for creators who want to turn a written scene or a single starting image into a short video. The Movelume workspace keeps both ways to begin in one place: choose text to video when the visual world begins in words, or image to video when a still already defines the composition, subject, or art direction. You can shape a four-to-thirty-second draft, choose 480p, 720p, or 1080p, and decide whether the clip should include generated audio. Text to video offers six frame shapes; image to video uses automatic framing.
That range makes Seedance 2.5 useful for more than a quick motion test. A short clip can test a camera move, reveal, or product beat before a team commits to a longer edit. A longer clip can hold a simple sequence, allow an action to settle, or give atmosphere time to register. The model is still a generation tool rather than a complete editor: it does not replace shot planning, continuity review, sound finishing, or the judgment needed to decide which take belongs in a final cut. Select Seedance 2.5 when you need a broad video draft rather than a specialized effect. The best results start with one clear shot objective and enough visual detail to guide movement without scripting every frame.
Use text to video to explore how a product might enter a frame, how light could travel across a material, or how a simple reveal might feel before the final shot is made. Describe the subject, setting, camera, movement, and lighting in one coherent scene. Keep logos, packaging text, and tiny labels out of the first attempt unless exact typography is essential; generated video is better for testing direction and rhythm than approving final brand artwork.
Use image to video when a key visual, storyboard frame, illustration, or product composition already carries the look you want. The starting image anchors subject placement and color, while the prompt explains what should move and how the camera should respond. A clean source with one readable focal point is easier to animate than a crowded collage. Leave visual room around subjects that need to move, and avoid asking the model to replace the entire scene at once.
Vertical and square frame shapes make the model useful for story beats, explainers, title backgrounds, mood transitions, and editorial cutaways. Start with one moment that can be understood without a paragraph of context. If the concept needs three locations, several costume changes, and precise dialogue, divide it into separate shots. Short, single-purpose clips are easier to evaluate, revise, and assemble into a sequence than one overloaded generation.
The controls below show what you can currently choose for Seedance 2.5 in Movelume. They cover the inputs, duration, resolution, frame shape, audio, and credits available on this page, so you can prepare a draft around options you can actually select.
Text to video or one starting image. Text to video begins from a scene description. Image to video accepts one first frame and uses the prompt to direct motion. Multiple references, a separate final frame, and a reference-video workspace are not part of this page.
4–30 seconds. Choose an explicit whole-second duration. Automatic duration is not shown, which keeps the expected length and credit estimate visible before you continue.
480p, 720p, or 1080p. Use 480p for lower-cost exploration, 720p for review, or 1080p for a more detailed output. Check the updated credit estimate before continuing.
Six text-to-video shapes · automatic image framing. Text to video offers 21:9, 16:9, 4:3, 1:1, 3:4, and 9:16. Image to video chooses framing automatically from your first frame. Keep subjects clear of edges.
A useful prompt gives the model a stable visual hierarchy. Put the main subject and action first, then describe the environment, camera behavior, light, and mood. Add constraints only when they change the visible result.
Begin with the person, object, creature, or environment the viewer should notice first. Pair it with a visible action: a cyclist leans into a wet corner, a glass bottle rotates on a stone plinth, or fog moves through a pine valley. Concrete verbs provide more useful direction than abstract claims such as cinematic, premium, or amazing. If several subjects matter, state their spatial relationship so the frame does not have to invent the hierarchy.
Choose a camera idea that supports the action: a slow push toward the subject, a lateral tracking shot, a locked tripod view, a handheld follow, or a gentle orbit. Avoid stacking contradictory moves such as a static frame, rapid crane, close orbit, and sweeping drone flight in the same sentence. Mention the lens or depth of field only when it is central to the look. The goal is a readable shot, not a list of filmmaking vocabulary.
Lighting, weather, material, and scale help movement feel connected to a world. Describe how light falls, what surfaces reflect, whether fabric catches wind, or how dust reacts to a step. These details are more actionable than a long style label. For image to video, keep the prompt compatible with the starting frame: ask the existing subject to move, the camera to travel, or the atmosphere to change instead of demanding a different cast and location.
Longer clips create more time for an action to develop, but they also create more frames in which identity, geometry, or background details may drift. Test the core motion at a shorter duration first when the concept is uncertain. Move to a longer clip after the subject, camera, and scene relationship are working. Review hands, faces, product edges, repeated patterns, signs, and contact points frame by frame before using a take in public work.
Plans for image and video
Annual plans are billed upfront. The same plan credits refresh every 30 days and expire at the end of that window; unused subscription credits do not roll over.
Explore the workspace and prepare a first request before choosing a paid plan.
For occasional image and video experiments with a predictable monthly ceiling.
Subscriptions renew automatically at the selected monthly or yearly rate until canceled. Cancel future renewals in your account.
For creators who iterate every week and need room for several directions.
Subscriptions renew automatically at the selected monthly or yearly rate until canceled. Cancel future renewals in your account.
For sustained, high-volume creation and access to the highest available output quality.
Subscriptions renew automatically at the selected monthly or yearly rate until canceled. Cancel future renewals in your account.
Prices are in USD. Subscriptions renew automatically; credit packs are one-time purchases.
Compare plansYes. Select text to video when you want to invent the scene from a written brief. Select image to video when one still already defines the subject, composition, or visual language. Image to video currently accepts one starting image. If you need several references to guide identity, wardrobe, products, or locations, this page does not support that input yet.
Choose 480p for early motion tests, prompt comparison, and lower-cost exploration. Choose 720p for reviewing the shot or 1080p when you need a more detailed output. Resolution alone does not fix a weak composition or inconsistent action, so improve the shot brief before paying for more pixels.
Resolution and duration affect the credit estimate. Text-to-video frame shape can also affect lower-resolution estimates. Image-to-video automatic framing and 1080p use a fixed rate for the selected resolution and duration. The workspace rounds the estimate to whole credits. Treat the shown value as the current estimate for that configuration, not a permanent promise about future pricing.
No. Duration gives an action more time, but it can also increase opportunities for visual drift. A concise four-to-eight-second shot is often the best place to test a camera move or product beat. Use a longer duration when the scene genuinely needs an opening, development, and settling moment, and when the shorter version has already shown stable subjects and geometry.
Not in the current Movelume Seedance 2.5 workspace. Image to video uses one first frame, and text to video uses no image. A separate final frame, multiple image references, and reference-to-video controls are not available. Choose the first frame that best establishes your subject and composition, then describe the intended ending in the motion prompt. An exact final composition still requires checking the delivered clip.
You can choose Seedance 2.5, switch between its supported workflows, set the visible controls, and keep the draft in your browser. Direct online generation is not available yet. The Generate action will not contact a video service or spend credits from this page while that capability is unavailable. You can still prepare a complete brief and compare another video model without losing your prompt.
Return to the generator above to configure Seedance 2.5, or open the model catalog to compare another video option. Your prompt and compatible settings can stay with the draft while you decide.
