AI and machine learning
OpenSearch provides machine learning (ML) algorithms, model integrations, model inference, agents, and related AI capabilities. Integrate ML models to generate embeddings or text, build agents that use models and tools to perform multi-step tasks, or run built-in algorithms to analyze your data.
Interactive demos
ML models
OpenSearch supports ML models that you can use to enhance search relevance through semantic understanding. You can either deploy models directly within your OpenSearch cluster or connect to models hosted on external platforms. These models can transform text into vector embeddings, enabling semantic search capabilities, or provide advanced features like text generation and question answering. For more information, see Integrating ML models.
Agents and tools
An agent uses a large language model (LLM) to plan and run multi-step tasks, such as answering questions about your data or generating queries from natural language. To perform these tasks, an agent calls tools, each of which runs a specific operation, such as running a search or listing indexes. For more information, see Agents and tools.
MCP servers and clients
Model Context Protocol (MCP) is an open standard that defines how AI applications discover and call tools. OpenSearch can act as an MCP client, so that its agents can call external tools, or as an MCP server, so that external AI applications can call OpenSearch tools. For more information, see MCP servers and clients.
Memory and context
Agents can retain information across conversations and control how much context they send to the LLM. For more information, see Memory and context.
Built-in algorithms for data analysis
OpenSearch includes built-in algorithms that analyze your data directly within your cluster, enabling tasks like anomaly detection, data clustering, and predictive analytics without requiring external ML models.