Tablestore integrates with various mainstream AI frameworks, platforms, and applications, providing high-performance data storage and retrieval for AI application development. This document describes the AI ecosystem integrations that Tablestore supports.
Agent open-source components
Tablestore for Agent Memory
The Agent Memory SDK is a framework built on Tablestore. It provides general-purpose memory storage and query capabilities for AI agent scenarios. It supports two primary scenarios: real-time memory storage and long-term semantic knowledge retrieval.
MCP protocol integration
The Model Context Protocol (MCP) is a standard protocol for managing the context of AI applications. Tablestore provides multiple MCP service implementations to support various application scenarios.
Tablestore MCP Server
A general-purpose Tablestore MCP service implementation:
Document: Quick Start for Tablestore MCP Service
Function Compute Function AI plugin: Tablestore MCP Server plugin
Tablestore OpenMemory MCP
A cross-platform AI memory service based on Mem0. It enables intelligent memory sharing and reuse across multiple applications through unified memory storage:
Function Compute Function AI plugin: Tablestore OpenMemory MCP plugin
AI framework integration
LlamaIndex
LlamaIndex is a widely used framework for developing large language model (LLM) applications. Tablestore fully supports its core storage components:
VectorStore: Stores high-dimensional vectors and supports similarity searches.
DocumentStore: Stores and manages original document content.
IndexStore: Manages and maintains various types of index metadata.
ChatStore: Saves and retrieves chat history.
KvStore: Provides general-purpose key-value storage.
LangChain
LangChain is a widely used framework for developing AI applications. Tablestore supports the following component:
VectorStore: Stores and retrieves vector embeddings.
LangChain4j
LangChain4j is the Java implementation of LangChain. Tablestore supports the following components:
EmbeddingStore: Manages embedding data from text vectorization.
MemStore: Manages conversation history and context.
LangEngine
LangEngine is an AI application development framework developed by Alibaba. Tablestore supports the following component:
VectorStore: Stores and retrieves vector embeddings.
AgentScope Runtime
AgentScope is an open-source, multi-agent platform from Alibaba. Tablestore supports the following components:
TablestoreRAGService: A retrieval-augmented generation (RAG) service that provides document retrieval and knowledge base query capabilities.
TablestoreMemoryService: An agent memory service that manages the long-term and short-term memory of agents.
AI application platform integration
Dify
Dify is an open-source LLM application development platform. Tablestore is integrated with the following component:
VectorStore: Provides vector retrieval capabilities for RAG applications.
Related documents:
PAI-RAG
PAI-RAG is a RAG application framework provided by Alibaba Cloud's Platform for AI (PAI). Tablestore is integrated with the following component:
VectorStore: Provides a high-performance vector retrieval service.
Related document:
Stable Diffusion
Stable Diffusion is a popular AI image generation model that supports scenarios such as text-to-image and image-to-image generation. Tablestore provides integration support for the following component:
Image management extension: Manages and organizes generated image resources. This extension is built into Alibaba Cloud Function Compute (FC) applications.