Boost Elasticsearch Semantic Search with Alibaba Cloud AI: Step‑by‑Step Guide

Author: Tomás Murúa (Elastic)

This article explains how to integrate Alibaba Cloud AI capabilities with Elasticsearch to improve semantic search relevance.

Steps

Create Elasticsearch mapping

Index data into Elasticsearch

Query data

Reward: complete answering questions

Configure Alibaba Cloud AI

Alibaba Cloud AI search provides semantic reranking and sparse embedding endpoints using models such as Qwen and DeepSeek‑R1.

Semantic reranking reorders results based on similarity between query and documents, while sparse embeddings highlight relevant information.

Obtain Alibaba Cloud API Key

Generate a valid API key from the Alibaba Cloud portal:

Access the Service Marketplace.

Navigate to the left‑hand menu API Keys .

Create a new API key.

Configure Alibaba Endpoints

Set up the sparse embedding endpoint:

PUT _inference/sparse_embedding/alibabacloud_ai_search_sparse
{
  "service": "alibabacloud-ai-search",
  "service_settings": {
    "api_key": "<api_key>",
    "service_id": "ops-text-sparse-embedding-001",
    "host": "default-j01.platform-cn-shanghai.opensearch.aliyuncs.com",
    "workspace": "default"
  }
}

Set up the rerank endpoint:

PUT _inference/rerank/alibabacloud_ai_search_rerank
{
  "service": "alibabacloud-ai-search",
  "service_settings": {
    "api_key": "<api_key>",
    "service_id": "ops-bge-reranker-larger",
    "host": "default-j01.platform-cn-shanghai.opensearch.aliyuncs.com",
    "workspace": "default"
  }
}

Create Elasticsearch Mapping

Define fields for semantic vectors and descriptions, using copy_to for hybrid search:

PUT arts
{
  "mappings": {
    "properties": {
      "semantic_description": {
        "type": "semantic_text",
        "inference_id": "alibabacloud_ai_search_sparse"
      },
      "description": {
        "type": "text",
        "copy_to": "semantic_description"
      }
    }
  }
}

Index Data

Bulk‑index a set of Agatha Christie novel and play descriptions:

POST arts/_bulk
{ "index": {} }
{ "description": "Black Coffee is a play by the British crime‑fiction author Agatha Christie..." }
{ "index": {} }
{ "description": "The Mousetrap is a murder mystery play by Agatha Christie..." }
{ "index": {} }
{ "description": "The Body in the Murder is a Miss Marple mystery novel published by Agatha Christie in 1942..." }
{ "index": {} }
{ "description": "Curtain: Poirot's Last Case is Agatha Christie's last published novel..." }
{ "index": {} }
{ "description": "Death on the Nile is Agatha Christie's most daring travel mystery novel..." }
{ "index": {} }
{ "description": "The Murder of Roger Ackroyd was Agatha Christie’s first book published by William Collins..." }

Semantic Search

Query the semantic_description field:

GET /arts/_search
{
  "_source": { "includes": ["description"] },
  "query": {
    "semantic": {
      "field": "semantic_description",
      "query": "Which novel was written by Agatha Christie?"
    }
  }
}

The response returns novel documents first, with plays appearing at the bottom.

Rerank Optimization

Use the rerank endpoint to improve ordering:

POST _inference/rerank/alibabacloud_ai_search_rerank
{
  "query": "Which novel was written by Agatha Christie?",
  "input": [
    "Black Coffee is a play ...",
    "The Mousetrap is a murder mystery play ...",
    "The Body in the Murder is a Miss Marple mystery novel ...",
    "Curtain: Poirot's Last Case ...",
    "Death on the Nile ...",
    "The Murder of Roger Ackroyd ..."
  ]
}

Combined Retrieval and Rerank

Execute semantic search and reranking in a single request:

POST /arts/_search
{
  "_source": { "includes": ["description"] },
  "retriever": {
    "text_similarity_reranker": {
      "retriever": { "standard": { "query": { "semantic": { "field": "semantic_description", "query": "Which novel was written by Agatha Christie?" } } } },
      "field": "description",
      "rank_window_size": 10,
      "inference_id": "alibabacloud_ai_search_rerank",
      "inference_text": "Which novel was written by Agatha Christie?"
    }
  }
}

Completion Endpoint for RAG

Create a completion endpoint using Alibaba Qwen or DeepSeek‑R1:

PUT _inference/completion/alibabacloud_ai_search_completion
{
  "service": "alibabacloud-ai-search",
  "service_settings": {
    "host": "default-j01.platform-cn-shanghai.opensearch.aliyuncs.com",
    "api_key": "<api_key>",
    "service_id": "ops-qwen-turbo",
    "workspace": "default"
  }
}

Send the retrieved documents and the question to obtain a concise answer from the LLM.

Conclusion

Integrating Alibaba Cloud AI with Elasticsearch enables seamless use of embedding, rerank, and completion models, enhancing search pipelines and moving toward a full RAG solution.