Skip to content

WAND-Create โ€‹

Updated

This is the API documentation for generating text, images, video, music, and 3D, and for analyzing media. Choose a task type first, then start from the request example of the model you will use.

Choosing a Task โ€‹

TaskDocumentation
Text generation and chatVOD Text / GPT, Gemini, GK, DeepSeek, GLM, Kimi, MiniMax, Hunyuan
Image generation and editingVOD Image, MPS Image, TokenHub Image
Video generationVOD Video, MPS Video
Music and sound generationVOD Music, MPS Music
Image and video understandingMedia Understanding / Vega-VL, Gemini

Call Paths โ€‹

CategoryEndpointAuthenticationResponse mode
VOD Image / Video / Music / 3Dvod.intl.tencentcloudapi.comTC3-HMAC-SHA256Query after the TaskId is accepted
MPS Image / Video / Music / 3Dmps.tencentcloudapi.comTC3-HMAC-SHA256Query after the TaskId is accepted
VOD Text / Media Understandingmmu.vod-qcloud.comVOD AIGC Bearer TokenComplete response or SSE
TokenHub Imagetokenhub-intl.tencentcloudmaas.comTokenHub API KeyPer-model response and query spec

For VOD generation, specify the output specification with OutputConfig; for MPS generation, with ExtraParameters. The model-specific extra options of VOD use ExtInfo, and MPS is passed to AdditionalParameters as a JSON string.

Vega Models โ€‹

TaskModelCall path
Image generation and editingWAND-Vega-Image / Lite / Flash / ProMPS / VOD / Tokenhub
Video generationWAND-Vega-Video / 1.5-pro / 1.0-pro / 1.0-liteMPS / VOD
Image and video understandingVega-VLVOD multimodal API

Querying Generation Results โ€‹

  1. Call the generation action and save TaskId.
  2. VOD checks DescribeTaskDetail for its Status=FINISH and ErrCode=0, and MPS checks the Status=DONE of the per-model query action.
  3. Download the file from the result URL, or check the storage specified in the generation request.

Text and Media Understanding do not require querying a generation task. Results are read from the HTTP response or SSE.

Document Structure โ€‹

Each API Reference explains authentication, common request and response formats, and the query and storage specifications. Model documents are organized in the order of basic information, request parameters, request examples, special settings, and response examples. The Image, Video, and Text menus are used to expand the model list.

Built with VitePress ยท Deployed on Tencent EdgeOne Pages