Seedance 1.0: Pro, Lite, Fast, Pricing and API (2026)

Seedance 1.0 is ByteDance’s first generation of AI video models: you give it a text prompt or a still image, and it returns a short video clip with realistic motion. It launched in 2025 and has since been followed by Seedance 2.0. The 1.0 family is still widely used because it is cheap, predictable and well understood. This guide explains the versions, what each one costs per clip, how to call it from code, and where it fits next to newer video models.

What is Seedance 1.0?

Seedance 1.0 is a family of generative video models built by ByteDance’s Seed team. Like most modern video generators, it is a diffusion model. That means it starts from random noise and refines it, step by step, into frames that match your prompt. Because it generates the whole clip together, rather than one frame at a time, motion stays coherent across the shot.

The family covers two jobs:

• Text-to-video (T2V): you describe a scene in words and the model invents everything in the frame.
• Image-to-video (I2V): you supply a picture and the model animates it. Your image becomes the first frame, so the product, character or composition stays exactly as you designed it.

Within the family there are three quality tiers. Lite is tuned for low cost. Pro is the flagship, with strong 1080p output, cinematic camera motion and multi-shot sequences, where one clip cuts between several angles of the same scene. Pro Fast is an image-to-video variant with a slimmer set of controls and flat per-clip pricing.

ByteDance offers Seedance inside its own consumer apps, such as Doubao. Developers usually reach it through an API. On Unicorn API, all five Seedance 1.0 variants sit behind the same key and the same credit balance as every other image, video, audio and chat model in the model catalog.

Seedance 1.0 Pro, Pro Fast and Lite: which version does what

Five versions of Seedance 1.0 are available on Unicorn API. The differences come down to input type, quality tier and which controls you get.

Version	Input	Resolutions priced	Pricing basis	Notable controls
Seedance 1.0 Lite Text to Video	Text	480p, 720p, 1080p	Per second	seed, aspect_ratio, camera_fixed
Seedance 1.0 Lite Image to Video	Image + text	480p, 720p, 1080p	Per second	seed, camera_fixed, end_image_url
Seedance 1.0 Pro Text to Video	Text	480p, 720p, 1080p	Per second	seed, aspect_ratio, camera_fixed
Seedance 1.0 Pro Image to Video	Image + text	480p, 720p, 1080p	Per second	seed, camera_fixed
Seedance 1.0 Pro Fast Image to Video	Image + text	720p, 1080p	Per clip (5 or 10 s)	prompt, duration, resolution only

Lite: drafts, volume and social clips

Lite is the version to reach for when you need many clips or when you are still working out the idea. At 480p it is the cheapest way to see whether a prompt produces the motion you want. The Lite image-to-video model also has a feature the Pro models lack here: end_image_url. You can give it both a start frame and an end frame, and the model fills in the motion between them. That is very useful for transitions, before-and-after shots and product reveals.

Pro: final output and multi-shot scenes

The Seedance 1.0 Pro AI video generator is where the family is strongest. It is built for 1080p, cinematic motion and clips that cut between shots while keeping the subject consistent. Use it for anything a client or audience will actually see. It costs more per second than Lite, so most teams draft on Lite and switch to Pro once the prompt is settled.

Pro Fast: simple, flat-priced image animation

Seedance 1.0 Pro Fast is image-to-video only and deliberately minimal. It has no seed and no camera lock, just prompt, image, duration, resolution and a content checker. It is priced per clip at fixed 5-second and 10-second lengths. At 1080p it works out at roughly half the price of Pro for the same length, which makes it a strong default for animating product shots or illustrations at scale.

Seedance 1.0 API pricing per clip

Unicorn API prices everything in credits. On a monthly plan a credit costs $0.01 (Starter $25, Creator $50, Studio $100 and Pro $200 per month). Top-up credits cost $0.015 each and never expire. Details are on the pricing page.

Lite and Pro are billed per second of output: the per-second rate times the duration you request, rounded up to whole credits. Pro Fast has a fixed price per clip.

Model	480p	720p	1080p
Lite (T2V or I2V), per second	2 credits	4.5 credits	10 credits
Pro (T2V or I2V), per second	2.8 credits	6 credits	14 credits
Pro Fast, 5-second clip	n/a	16 credits	36 credits
Pro Fast, 10-second clip	n/a	36 credits	72 credits

What a 5-second clip costs

Model and resolution	Credits	On a plan ($0.01)	As top-up ($0.015)
Lite 480p	10	$0.10	$0.15
Lite 720p	23 (22.5 rounded up)	$0.23	$0.345
Lite 1080p	50	$0.50	$0.75
Pro 480p	14	$0.14	$0.21
Pro 720p	30	$0.30	$0.45
Pro 1080p	70	$0.70	$1.05
Pro Fast 720p	16	$0.16	$0.24
Pro Fast 1080p	36	$0.36	$0.54

Two things stand out. First, Pro Fast at 720p is cheaper than Lite at 720p for a 5-second clip (16 credits against 23). If you have a start image, Pro Fast is often the best value in the whole family. Second, resolution is the biggest cost lever. Going from 720p to 1080p on Pro more than doubles the price, so only render 1080p when the output really needs it.

A worked budget

Say you are making three 5-second product teasers. You draft 20 variations on Lite text-to-video at 480p (20 × 10 = 200 credits). You then render your three best prompts on Pro at 1080p (3 × 70 = 210 credits). The total is 410 credits, or $4.10 from a monthly plan. A Starter plan covers that with plenty to spare. You never have to guess, either: send the request to POST /v1/models/price and you get the exact cost before anything runs.

How to generate a video with the Seedance 1.0 API

Every video model on Unicorn API uses the same asynchronous job flow. Video takes longer than a normal HTTP request should stay open, so you submit a job, get an ID back, and poll until it finishes.

1. Create a key at /keys and export it as UNICORN_API_KEY.
2. Optionally check the cost with POST /v1/models/price.
3. Submit the job with POST /v1/jobs, passing the model ID and an input object.
4. Poll GET /v1/jobs/{id} until status is "succeeded".
5. Download the video from outputs[0].url.

Quick test with curl

curl -X POST https://api.unicornapi.net/v1/jobs \
  -H "Authorization: Bearer $UNICORN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "bytedance-v1-pro-fast-image-to-video",
    "input": {
      "prompt": "A slow drone shot over a rocky coastline at sunset, waves breaking",
      "image_url": "https://example.com/image.png",
      "duration": 5,
      "resolution": "720p"
    }
  }'

The response includes a job ID. Poll it with:

curl https://api.unicornapi.net/v1/jobs/JOB_ID \
  -H "Authorization: Bearer $UNICORN_API_KEY"

Complete Python script

This script checks the price, submits a Seedance 1.0 Pro text-to-video job, waits for it and saves the MP4. It uses only the requests library.

import os
import time
import requests

API = "https://api.unicornapi.net"
HEADERS = {
    "Authorization": f"Bearer {os.environ['UNICORN_API_KEY']}",
    "Content-Type": "application/json",
}

body = {
    "model": "bytedance-v1-pro-text-to-video",
    "input": {
        "prompt": (
            "A slow drone shot over a rocky coastline at sunset, "
            "waves breaking, warm golden light, cinematic"
        ),
        "duration": 5,
        "resolution": "1080p",
        "aspect_ratio": "16:9",
        "seed": 42,
    },
}

# 1. Check the exact price before spending credits
price = requests.post(f"{API}/v1/models/price", headers=HEADERS, json=body)
price.raise_for_status()
print("Price:", price.json())

# 2. Submit the job
resp = requests.post(f"{API}/v1/jobs", headers=HEADERS, json=body)
resp.raise_for_status()
job = resp.json()
job_id = job["id"]
print("Submitted job", job_id)

# 3. Poll until the job finishes
while True:
    job = requests.get(f"{API}/v1/jobs/{job_id}", headers=HEADERS).json()
    status = job["status"]
    if status == "succeeded":
        break
    if status in ("failed", "canceled", "error"):
        raise RuntimeError(f"Job ended with status {status}: {job}")
    time.sleep(5)

# 4. Download the video
url = job["outputs"][0]["url"]
video = requests.get(url)
video.raise_for_status()
with open("seedance.mp4", "wb") as f:
    f.write(video.content)
print("Saved seedance.mp4 from", url)

To switch versions, change the model string. Image-to-video models need an image_url, and the Lite image model also accepts end_image_url. Pro Fast does not take seed or aspect_ratio, so remove them when you use it. The full field list for each model is in the API documentation. If you would rather try prompts before writing code, the video playground runs the same models in your browser. If you work in Cursor or Claude Code, the MCP setup lets your agent call the API directly.

Prompting tips that work with Seedance 1.0

Seedance 1.0 responds best to prompts written like a short shot description: subject, action, camera, light. A few habits make a big difference.

• Lead with the subject and the action. “A ceramic mug on a wooden table, steam rising slowly” works better than a list of adjectives. The model needs to know what moves.
• Name the camera move. Phrases such as “slow dolly in”, “aerial drone shot”, “handheld tracking shot” or “static tripod shot” steer the motion. If you want no camera movement at all, set camera_fixed to true on Lite or Pro instead of relying on the prompt.
• Describe light and mood once. “Golden hour”, “overcast soft light” or “neon reflections on wet asphalt” give the model a clear look without overloading the prompt.
• For image-to-video, describe motion, not the picture. The model can already see your image. Spend the prompt on what should happen: “the model turns her head toward the camera and smiles” rather than restating her outfit.
• For multi-shot on Pro, write the cuts. Describe the shots in order, for example “wide shot of a runner on a bridge, then close-up of her shoes hitting the pavement”. Keep the subject the same so the model holds it consistent across cuts.
• Fix the seed while iterating. A seed is a number that controls the random starting noise. Keeping it the same means a prompt change is the only thing that changes, so you can see what your edit actually did.

Common mistakes with Seedance 1.0

Most wasted credits come from a handful of avoidable errors.

1. Drafting at 1080p on Pro. A 5-second Pro 1080p clip costs 70 credits. The same prompt on Lite at 480p costs 10. Settle the motion cheaply, then upscale your choice to Pro.
2. Forgetting the per-second multiplier. Lite and Pro prices are per second. Doubling duration doubles the bill. Pro Fast is the exception, with flat prices for 5 and 10 seconds.
3. Sending text-only requests to image-to-video models. The I2V variants expect image_url. If you have no image, use a T2V model.
4. Passing fields a model does not accept. Pro Fast has no seed, and image-to-video models take their shape from your image rather than aspect_ratio. Check the field list on each model page.
5. Using low-resolution or cluttered start images. The first frame sets the quality ceiling. A clean, sharp, well-lit image gives far better motion than a compressed thumbnail.
6. Polling too aggressively. Checking every few seconds is plenty. Hammering the status endpoint does not make the job finish sooner.
7. Treating the URL as permanent storage. Download the file as soon as the job succeeds and store it yourself.

Seedance 1.0 vs newer video models

Seedance 1.0 is no longer the newest video model, and that is partly why it is useful: it is cheap and its behaviour is well known. Here is how it compares on price and features with some newer models available under the same key.

Model	Input	5 s at 720p	5 s at 1080p	Stand-out feature
Seedance 1.0 Lite	Text or image	23 credits	50 credits	Start and end frame (I2V)
Seedance 1.0 Pro	Text or image	30 credits	70 credits	Multi-shot, cinematic motion
Seedance 1.0 Pro Fast	Image	16 credits	36 credits	Flat per-clip pricing
Grok Imagine Video 1.5	Image	23 credits (22.5 rounded up)	40 credits	Any length from 1 to 15 s
Grok Imagine Video 1.0	Text or image	n/a (6 s minimum)	n/a (6 s minimum)	Clips from 6 up to 30 s
PixVerse V6 Max Image To Video	Image	44 credits (no audio)	88 credits (no audio)	Optional generated audio
Kling 4.0 Keyframe To Video	Keyframes + text	143 credits (142.5 rounded up)	213 credits (212.5 rounded up)	Several keyframes per clip

The trade-offs break down like this:

• Lowest cost for animating an image: Seedance 1.0 Pro Fast at 720p is the cheapest 5-second option in this table.
• Odd lengths or long clips: Grok Imagine 1.5 lets you pick any whole number of seconds from 1 to 15, and Grok Imagine 1.0 goes up to 30 seconds. Seedance Pro Fast is fixed at 5 or 10. The Grok Imagine guide covers both.
• Sound in the clip: PixVerse V6 Max can generate audio alongside the video. Seedance 1.0 output is silent, so you would add music or voice in editing.
• Precise control over the path of a shot: Kling 4.0 Keyframe accepts a first frame, a last frame and extra keyframes in between. It costs much more, but it gives you the most control. For a simple start-to-end transition, Seedance Lite’s end_image_url does a similar job for far less.

Because all of these share one API key, one job flow and one credit balance, switching models is a one-line change to the model field. A practical setup is to default to Seedance 1.0 for volume work and route special cases, such as audio, long clips or keyframe control, to the model built for them.

Is Seedance 1.0 open source, and is there an app?

Seedance 1.0 is not open source. ByteDance has not published the model weights, so you cannot download it and run it on your own GPUs. Every legitimate way to use it is hosted: ByteDance’s own apps, or an API provider.

That also answers the “Seedance 1.0 app download” question. There is no standalone official Seedance app. ByteDance offers its Seedance models inside its own consumer products, such as Doubao. Be wary of sites offering a “Seedance 1.0 desktop download” or model files. At best these are wrappers around someone else’s API, and at worst they are malware.

For developers, the hosted route is the better choice anyway. You pay per clip, you know the price in advance, and you can move to a newer model with one line of code when your needs change. Create a key at /keys, run the Python script above with a 480p Lite draft, and you will have your first Seedance 1.0 clip for about 10 credits.