All Products
Search
Document Center

OpenSearch:Multimodal embedding

Last Updated:Aug 05, 2025

AI Search Open Platform supports calling multimodal embedding services through SDK.

Prerequisites

  • You have activated the AI Search Open Platform service. For more information, see Activate the service.

  • You have completed identity authentication using API-KEY. For authentication information, see Obtain API-KEY.

  • To call the multimodal embedding service, you need to update the SDK version to 2.1.0 or later.

    pip install --upgrade alibabacloud_searchplat20240529

Parameters

from alibabacloud_tea_openapi.models import Config
from alibabacloud_searchplat20240529.client import Client
from alibabacloud_searchplat20240529.models import GetMultiModalEmbeddingRequest, GetMultiModalEmbeddingRequestInput

if __name__ == '__main__':
    # token configuration, endpoint configuration
    config = Config(bearer_token="Replace with your API-KEY",
                    # endpoint: Configure the unified request entry point, remove http://
                    endpoint="Replace with API endpoint",
                    # Specify the request protocol. Valid values: HTTPS and HTTP.
                    protocol="http")
    client = Client(config=config)
 
    # --------------- Request body parameters ---------------
    request = GetMultiModalEmbeddingRequest()
    request.from_map({"input":[{"text":"Science and technology are the primary productive forces"}]})

    # default: Replace with your workspace name, ops-m2-encoder: service ID, can be replaced with other supported service IDs
    response = client.get_multi_modal_embedding("default", "ops-m2-encoder", request)
    print(response)