Quick start

Use the Lingjuta Model Platform to access text, image, and video models.

Before you begin

Lingjuta provides a unified entry point for model integration. This guide walks you through preparing your key, choosing a model, and making your first video task request.

What you need

A platform account, an API key, and a server environment that can send HTTPS requests.

Step 1: Get an API key

1. Open API key management.
2. Sign in and create an API key.
3. Store the key securely and read it from an environment variable on your server.

Do not include keys in browser code, public repositories, or screenshots. See Authentication for details.

Step 2: Choose a model

Type	Integration protocol	Notes
Text	OpenAI-compatible format	Choose an available model from the model catalog; see Text integration
Image	Depends on the model	Input and output structures may differ by model; see Image integration
Video	Seedance native protocol	Submit an asynchronous task, then query it by task ID; see Create a video task

Refer to the model catalog and your account configuration for available models and permissions.

Step 3: Send a request

The example below uses the Seedance native protocol, whose API definition has been reviewed. Set API_BASE_URL to the API service root provided by the platform, without a /v1 or /api/v3 suffix.

export API_BASE_URL='https://YOUR_API_HOST'
export LINGJUTA_API_KEY='YOUR_API_KEY'

curl --fail-with-body "${API_BASE_URL}/api/v3/contents/generations/tasks" \
  -H "Authorization: Bearer ${LINGJUTA_API_KEY}" \
  -H 'Content-Type: application/json' \
  -d '{
    "model": "seedance-2.0-mini",
    "content": [{ "type": "text", "text": "A cat runs along the beach as the camera slowly follows" }]
  }'

A successful submission returns a task ID:

{
  "id": "00000000-0000-4000-8000-000000000001"
}

This means the task has been accepted. Video generation takes time; continue by querying the task status.

Step 4: Query the asynchronous task

curl --fail-with-body \
  "${API_BASE_URL}/api/v3/contents/generations/tasks/YOUR_TASK_ID" \
  -H "Authorization: Bearer ${LINGJUTA_API_KEY}"

queued and running mean the task is still in progress. succeeded means the video is ready at content.video_url. failed, cancelled, and expired are the other terminal states.

Use a reasonable polling interval and a maximum wait time, and stop at a terminal state. See Query a video task.

Next steps

• Video walkthrough: explore the full request flow and task lifecycle.
• Integration guide: handle timeouts, duplicate submissions, and errors.
• FAQ: troubleshoot keys, model permissions, and task outputs.