dreamina-3-0/text-to-video
Generate lifelike motion visuals fast with Dreamina 3.0 for designers.
Seedance 2.5 First & Last Frame is ByteDance Seed's early-access image-to-video mode for bridging two stills into one continuous clip. You supply a start image and an end image; the model invents the motion, camera path, and pacing that connect them while keeping both endpoint looks intact.
Output format: Video / up to ~30 seconds / 720p / aspect ratios from square to ultrawide
Use these fields to steer generation reliably.
| Parameter | Required | Type | Default | Range / Options | Description |
|---|---|---|---|---|---|
| prompt* | Yes (*) | string | — | — | Text describing the transition, action, and camera motion. |
| first_frame* | Yes (*) | string (image URL) | — | — | Starting keyframe image. |
| last_frame* | Yes (*) | string (image URL) | — | — | Ending keyframe image. |
| duration | No | integer | 5 | 4–30 | Output length in seconds. |
| aspect_ratio | No | string | 16:9 | 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, 9:21 | Output frame shape. |
| seed | No | integer | random | — | Fixed seed for reproducible runs; leave empty for random. |
Billing is based on the length of the generated clip.
| Billing unit | Rate |
|---|---|
| Per second of generated video | $0.70 |
Estimated cost examples
| Duration | Approx. cost |
|---|---|
| 5 s (default) | ~$3.50 |
| 10 s | ~$7.00 |
| 15 s | ~$10.50 |
| 30 s | ~$21.00 |
1) Open the model on RunComfy and prepare two stills that clearly mark the start and end of the shot.
2) Upload First Frame Image and Last Frame Image as the opening and closing keyframes.
3) Write a prompt for what happens between those frames — action, camera move, lighting, and mood.
4) Set Aspect Ratio to match delivery (for example 16:9 landscape or 9:16 vertical).
5) Choose Duration from 4–30 seconds; start shorter while iterating, then extend.
6) Optionally set a Seed to reproduce a strong take; leave empty for random.
7) Generate, review, then refine the prompt or swap an endpoint image if mid motion drifts.
8) For API use on RunComfy, send the same parameters; results appear in your job history.
On RunComfy you can iterate in the browser and reuse the same settings over HTTP API.
Generate lifelike motion visuals fast with Dreamina 3.0 for designers.
Master complex motion, physics, and cinematic effects.
AI-driven footage transformation with stable motion and design control
Create rapid high-quality video drafts with precise style and speed
HappyHorse 1.0 with native 1080p output, cinematic motion, and multi-shot consistency.
Features smooth scene transitions, natural cuts, and consistent motion.
Seedance 2.5 First & Last Frame is built for bridging two stills into one continuous clip. You upload a start image and an end image, then describe the motion between them so editors, product teams, and storyboard artists can link known keyframes without hard cuts.
Single-image modes only lock the opening look. Seedance 2.5 First & Last Frame also locks the closing still, so the model has to invent a path that arrives at your chosen end frame while keeping both endpoint identities recognizable.
It fits scene-to-scene bridges, pack-to-in-use product reveals, before-to-after transformations, and story beats where you already have two boards and need natural camera motion between them. Clear, related endpoint stills usually work better than unrelated images.
Keep subject scale and lighting similar across the two stills, and write a prompt that names camera moves plus a midpoint action. Seedance 2.5 First & Last Frame follows those cues more reliably than a prompt that only says "morph" or "transition."
You must provide a prompt and exactly two images in order: first frame, then last frame. Aspect ratio options include 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, and 9:21. Duration is a whole number of seconds from 4 to 30, with 5 as the default, and output is 720p.
Yes. Preserving the visual identity of the start and end frames is a core goal of Seedance 2.5 First & Last Frame. Results still depend on how related the two stills are and how clearly the prompt describes the journey between them.
Yes. Prototype in the RunComfy Web UI, then call the same Seedance 2.5 First & Last Frame model through the RunComfy API with the same parameters for production or automation. You do not need to host or scale the model yourself.
Generations consume usd / credits based on the generated video duration at $0.70 per second. Check the Generation section on this page for live pricing when you run Seedance 2.5 First & Last Frame.
RunComfy is the premier ComfyUI platform, offering ComfyUI online environment and services, along with ComfyUI workflows featuring stunning visuals. RunComfy also provides AI Models, enabling artists to harness the latest AI tools to create incredible art.
