Overview
The Elastic Agent Builder MCP server is the recommended approach for Elasticsearch 9.2+ deployments and Serverless projects. It offers full access to built-in and custom tools, enabling AI agents to connect to Elasticsearch data through natural language conversations with advanced retrieval and analysis capabilities.
Features
Agent Builder Capabilities
- Built-in Tools: Pre-configured tools for common operations
- Custom Tools: Create specialized tools for your use cases
- Full Tool Access: Complete access to Elasticsearch functionality
- Natural Language: Conversational interface to your data
Search Operations
- Lexical Search: Traditional keyword-based search
- Semantic Search: Vector-based similarity search
- Hybrid Search: Combine lexical and semantic for best results
- Complex Queries: Multi-field, nested, and aggregated searches
Data Analysis
- Aggregations: Analyze data patterns and trends
- Time Series: Work with time-based data
- Geospatial: Location-based queries and analysis
- Machine Learning: Access Elasticsearch ML features
Advanced RAG
- Hybrid Retrieval: Lexical + semantic search
- Reranking: Improve result relevance
- Context Enhancement: Enrich results with metadata
- Citation Tracking: Link answers to source documents
Search Engine
- Distributed search and analytics
- Real-time indexing and search
- Full-text search capabilities
- RESTful API
- JSON document storage
- Schema-free flexibility
- Automatic field mapping
- Nested and complex data types
Analytics
- Aggregation framework
- Time-series analytics
- Geospatial analytics
- Machine learning integrations
Use Cases
Enterprise Search
- Search across documents and databases
- Semantic understanding of queries
- Personalized search results
- Real-time index updates
Log Analytics
- Centralized log aggregation
- Real-time log search
- Pattern detection and alerts
- Performance monitoring
RAG Applications
- Build knowledge bases for AI
- Retrieve relevant context for LLMs
- Combine structured and unstructured data
- Explainable AI with source citations
Business Intelligence
- Real-time dashboards
- Data exploration and discovery
- Trend analysis and forecasting
- Custom analytics workflows
Version Compatibility
Elasticsearch 9.2+
- Agent Builder MCP (recommended)
- Full feature set
- Best performance and support
Earlier Versions
- elastic/mcp-server-elasticsearch
- Limited tool set
- Basic functionality
Integration
Works with AI assistants supporting MCP. Requires Elasticsearch 9.2+ or Elasticsearch Serverless. Configure with Elasticsearch credentials and endpoint URL.
Pricing
Elasticsearch offers various deployment options:
- Elastic Cloud: Managed service with usage-based pricing
- Self-Managed: Free basic license, paid for advanced features
- Serverless: Pay per search and storage
MCP server included with all Elasticsearch deployments.
