HiAPI Seedance Python SDK

Python SDK for Seedance 2.5 API, Seedance text-to-video, Seedance image-to-video, and Seedance video editing through the HiAPI unified task API.

Use this SDK when you want a small Python client focused on AI video generation:

• Seedance 2.5 API access through HiAPI's /v1/tasks endpoint.
• Text-to-video with client.text_to_video(...).
• Image-to-video with client.image_to_video(...).
• Video editing with client.video_edit(...).
• Async task polling built in: submit, poll, and read the output URL.
• Zero runtime dependencies: standard library only.

Install

pip install hiapi-seedance

Requires Python 3.8+.

Quick Start

from hiapi_seedance import Seedance

client = Seedance()  # uses HIAPI_API_KEY

task = client.text_to_video(
    prompt="A 30-second cinematic one-take product film, golden hour, slow dolly in",
    duration=30,
    aspect_ratio="16:9",
    resolution="1080p",
)

print(task.output[0].url)

Set your API key:

export HIAPI_API_KEY="sk-your-hiapi-key"

Get a key at HiAPI.

Text-to-Video

from hiapi_seedance import Seedance

client = Seedance(api_key="sk-...")

task = client.text_to_video(
    prompt=(
        "A one-take AI video prompt: a glass perfume bottle rotates on wet black "
        "stone while sunrise light moves across the surface, macro texture, no text."
    ),
    duration=15,
    aspect_ratio="16:9",
    resolution="1080p",
)

print(task.status, task.output[0].url)

Image-to-Video

task = client.image_to_video(
    prompt="Animate the reference image with a slow push-in, drifting fabric, and warm rim light.",
    image_urls=["https://example.com/first-frame.png"],
    duration=8,
    aspect_ratio="9:16",
)

Video Editing

task = client.video_edit(
    prompt="Keep the original camera movement and timing. Add a blue-white energy bow in the actor's hands.",
    video_urls=["https://example.com/source.mp4"],
    image_urls=["https://example.com/prop-reference.png"],
)

Return Immediately Instead of Waiting

created = client.text_to_video(
    prompt="A fashion film in a desert gallery",
    duration=30,
    wait=False,
)

task = client.wait(created.task_id, poll_interval=3, timeout=900)
print(task.output[0].url)

Raw HiAPI Task Shape

The SDK sends requests to:

POST https://api.hiapi.ai/v1/tasks
Authorization: Bearer $HIAPI_API_KEY
Content-Type: application/json

{
  "model": "seedance-2-5",
  "input": {
    "prompt": "A cinematic Seedance 2.5 prompt",
    "duration": 30,
    "aspect_ratio": "16:9",
    "resolution": "1080p"
  }
}

You can pass additional model fields as keyword arguments:

task = client.text_to_video(
    prompt="A multilingual typography loop",
    duration=15,
    web_search=False,
    generate_audio=True,
)

Prompt Library

Need production prompts and preview examples? Use the companion prompt library:

• Seedance 2.5 prompt library
• Seedance 2.0 prompt library

Model Slug

The default model is seedance-2-5. You can override it per client or per call:

client = Seedance(model="seedance-2-5")
task = client.text_to_video(prompt="...", model="seedance-2-0")

Use the fallback model while waiting for Seedance 2.5 availability on your account.

SEO Keywords

Seedance 2.5 API, Seedance Python SDK, Seedance API Python, Seedance text to video API, Seedance image to video API, AI video generation API, text-to-video Python, image-to-video Python, HiAPI Seedance API.

License

MIT