Tablestore is a Serverless table storage service designed for massive structured data. It supports four major data models: WideColumn, AI (Agent Memory), TimeSeries, and Timeline, making it suitable for scenarios such as AI data, metadata, messaging data, and spatio-temporal data. It provides capabilities such as secondary indexes, vector search, and SQL queries, and can be seamlessly integrated with compute ecosystems including MaxCompute, Flink, Spark, and Presto. It is also compatible with mainstream AI frameworks such as Dify, LangChain, and LlamaIndex. For AI intelligent Q&A, Tablestore offers an MCP-based intelligent agent architecture, and for IoT requirements, it provides the IoTstore solution to full-scenario data storage and intelligent analysis.
Product advantages
Multi-model and ecosystem compatibility: A cloud-native Serverless distributed NoSQL database that provides one-stop multi-model data storage capabilities. It is deeply integrated with the Alibaba Cloud compute ecosystem and open-source AI ecosystems to meet the data storage and processing requirements of complex business scenarios.
Serverless: Built on a pure Serverless architecture, it provides a fully managed distributed storage service. You do not need to worry about the provisioning, maintenance, and scaling of underlying hardware resources, and can focus on business logic development. The system automatically scales in and out based on business load.
High performance: Powered by a self-developed distributed storage engine, Tablestore delivers stable performance in read and write latency, concurrent throughput, and data retrieval, meeting the performance requirements of online services and big data analytics.
For more information, see What is Tablestore?.
Memory and knowledge storage services
Knowledge and memory storage are Serverless data services provided by Tablestore for AI agents. The knowledge storage service converts business documents into searchable knowledge to provide context for RAG applications. The memory storage service stores and recalls user preferences, historical facts, and session information to help agents maintain context across sessions.
The two services can be used independently or together. For example, an intelligent customer service bot can use a knowledge base to store product manuals and FAQs, and use a memory store to store user preferences and interaction history. When answering questions, it can obtain both authoritative materials and personalized context at the same time.
Typical application scenarios:
Multi-agent knowledge sharing
Agent data graph construction
User profile accumulation
Enterprise knowledge Q&A systems
RAG pipeline integration
For more information, see Memory and knowledge storage services.