GPTProto Python SDK

Typed Python client for the GPTProto unified asynchronous media API. The generated transport and models are wrapped by GptprotoClient for unified create/get/run polling.

Use

pip install gptproto
export GPTPROTO_API_KEY=your_key

from gptproto import GptprotoClient
from gptproto.models import UnifiedVideoRequest, VideoResolution

client = GptprotoClient()
result = client.run(
    "video",
    UnifiedVideoRequest(
        model="kwaivgi/kling-v2.5",
        prompt="a cat dancing",
        duration=5,
        resolution=VideoResolution.RESOLUTION_1080P,
    ),
)
print(result.unsigned_urls)

Task kinds are video, image, speech, voice-clone, lip-sync, 3d, and image-tool. run() returns on completed, raises TaskFailedError for failed, cancelled, or expired, and raises TaskTimeoutError on timeout.

For local development:

python3.11 -m venv .venv
source .venv/bin/activate
pip install -e .
python examples/demo.py

Use the repository-level ../generate.sh to regenerate without overwriting the hand-written run.py and public __init__.py.