Overview

OpenSearch is a one-stop platform to develop intelligent search services. OpenSearch was built based on the large-scale distributed search engine developed by Alibaba. OpenSearch serves more than 500 business cases in Alibaba Group and thousands of Alibaba Cloud customers. OpenSearch helps develop search services in different search scenarios, including e-commerce, O2O, multimedia, the content industry, communities and forums, and big data query in enterprises. OpenSearch helps you develop high quality, maintenance-free, and high performance intelligent search services to provide your users with high search efficiency and accuracy.

Benefits

Conversational search service

Build a dedicated conversational search system for enterprises based on the built-in large language model (LLM) and the business data of enterprises.

Search Practices in Multiple Industries

OpenSearch integrates the industry-specific knowledge graphs of Alibaba DAMO Academy, which helps develop industry-specific and intelligent search services in industry-specific search scenarios.

Search Algorithm Center

You can manage versions of algorithm models, iterate algorithm models, and train algorithm models based on data backflow in the search algorithm center.

Multimodal Retrieval

OpenSearch delivers high performance and reduces costs if you build vector indexes for a large amount of unstructured data, such as text, images, and behavioral data, and retrieve data based on the vector indexes.

Vector Search

Deeply integrated with Damo Institute's self-developed vector search engine Proxima to efficiently build vector search services, providing multimodal search for text/image/audio/video, etc.

Data Update within Millisecond

Thousand QPS millisecond query response.

Features

Conversational search service

OpenSearch LLM-Based Conversational Search Edition helps enterprises quickly build a dedicated conversational search system based on the built-in LLM. Dedicated conversational search services for enterprises Build a dedicated LLM and conversational search system based on the business data of enterprises to ensure security, accuracy, high quality, and reliability. Multimodal search results Support multimodal search results such as answers, URLs, and images to meet specific search requirements of users. Streaming output Support streaming output that can provide faster conversation response and better user experience.

Develop Industry-Specific and Intelligent Search Services

OpenSearch adopts the Natural Language Processing (NLP) technology developed by Alibaba DAMO Academy to implement intelligent semantic understanding on search queries. You can customize sort policies for your intelligent search services based on your business needs. Visualized Service Development Provides visualized interfaces throughout the development of search services and comes with various templates for industries, such as e-commerce. You can implement all the features of search services without writing code. Query Analysis Adopts the NLP technology developed by Alibaba DAMO Academy to implement intelligent semantic understanding on search queries and provides the following capabilities based on intelligent semantic understanding: analysis, synonym configuration, spelling correction, stop word filter, term weight analysis, named entity recognition (NER), and category prediction. Custom Sort Policies Supports rough sort and fine sort and allows you to customize sort policies based on your business needs. You can use Cava-based sort plug-ins for exclusive applications to implement complex sorting.

Develop Search Guidance Services

OpenSearch provides common peripheral features to guide searches, such as drop-down suggestions, top searches, and hints. OpenSearch also provides algorithm services, such as popularity models and category prediction models. You can use the algorithm services to customize features and improve the quality of intelligent search services. Drop-Down Suggestions Customizes drop-down suggestions models based on the data pushed to applications, recommends candidate search queries to guide searches and provides extended features, such as blacklists, whitelists, and intelligent sorting Top Searches and Hints Customizes top search models and hint models based on the user search behavior, generates top searches and hints to guide searches, and provides extended features, such as blacklists, whitelists, and top search ranking Algorithm Services Provides intelligent search algorithm services, such as popularity models and category prediction models, to optimize sorting and allows you to customize algorithms for different features of a search service

Develop High-Performance Data Search Services

OpenSearch accelerates database queries based on the core search engine that Alibaba Cloud has developed for years. Different Methods for Importing the Cloud and On-Premises Data Sources Allows you to connect OpenSearch to ApsaraDB RDS, MaxCompute, and PolarDB data sources on Alibaba Cloud with a few simple steps and upload on-premises data on demand by using the OpenSearch API or OpenSearch SDKs. Highly Efficient Data Update After a number of documents are pushed to a standard application, 99% of the documents can be searched in one second. After a number of documents are pushed to the primary table of an advanced application, 90% of the documents can be searched in ten seconds. Database Query Acceleration with High Performance Accelerates database queries based on the core search engine developed by Alibaba Cloud and delivers four times the performance of open-source search engines

Vector Search Service

OpenSearch Vector Search Edition is designed for vector scenarios and helps enterprises build high-performance, low-cost, easy-to-use, and powerful vector search services. High cost performance The minimum specification can support 1 billion 128-dimensional vectors, or thousand QPS millisecond query response; support data compression, fine index structure design, non-all-memory loading method, effectively reducing memory overhead. Rich vector capability Support HNSW, QC and other vector retrieval algorithms; support mixed retrieval between label, text inverted index and vector index; support filtering by expression, with the ability to filter while retrieving. Real-time writing of vector data It supports real-time writing of vector data at high qps, real-time dynamic updating of indexes on-the-fly, and fast adjustment of cluster data storage limit.