Build and deploy custom AI agents fast with Elastic Agent Builder
Elastic Agent Builder lets you quickly create custom agents that utilize all your data through skills, tools, and prompts — powered by Elasticsearch's hybrid search.
Ship agents grounded in your data in minutes
Chat with your data instantly, and build reliable agents with context engineering primitives that deliver all the relevant data and tools.
With the built-in Elastic AI Agent, users can securely query and analyze data in Elasticsearch indices in natural language.
Connect all your apps, agents, and data
Built-in MCP and A2A servers let external apps and agents call the tools and agents you build, with a REST API for everything else. Agent Builder connects to your data wherever it lives and works with any model, Elastic-managed or your own.
Extend, customize, and manage agents with the Elastic Agent Builder API
Each step is a single REST call: register a tool, compose it into an agent, then converse.
POST /api/agent_builder/tools
{
"id": "find_client_exposure_to_negative_news",
"type": "esql",
"description": "Finds client portfolio exposure to negative news",
"configuration": { "query": "ES|QL query here" },
"params": { "time_duration": { "type": "keyword" } }
}Take the next steps
Start building custom agents with tutorials, workshops, and technical blogs.
Frequently asked questions
Elastic Agent Builder is a conversational environment for building AI agents that answer questions and take actions over your data using natural language. Agents combine large language model (LLM) reasoning and context engineering with built-in and custom tools that query your indices, so responses stay grounded in your data. It's available in Elastic Cloud.
Use Elastic Agent Builder to create conversational AI agents, skills, and tools that retrieve, reason, and act on your data. Build agents that investigate incidents and find root cause, run threat hunts, create dashboards, or power knowledge and support assistants, all grounded in your data.
Elastic Agent Builder is model-agnostic. On Elastic Cloud Hosted and Elastic Cloud Serverless, you can use Elastic-managed models through the Elastic Inference Service with zero setup, including Jina for embeddings and reranking. You can also bring your own API key to connect external providers such as Anthropic, OpenAI, Amazon Bedrock, Azure AI Studio, and Google AI.
Elastic Agent Builder is built for data-intensive use cases, with native retrieval over your Elasticsearch indices through ES|QL, dense vector and semantic search, and BM25. General-purpose frameworks like LangGraph and Amazon Bedrock Agents leave retrieval for you to add and configure. If you're orchestrating with one of those frameworks, you can add Agent Builder as the specialist knowledge agent it calls over A2A or MCP.
Yes. Elastic Agent Builder supports both MCP and A2A. It ships with a built-in MCP (Model Context Protocol) server that makes your tools and agents available to MCP-enabled apps like Claude Desktop and Cursor, and an A2A (agent-to-agent) server for collaborating with external agents. A REST API is also available for custom integrations.
Elastic Agent Builder is free for the first 1,000 executions on Elasticsearch Serverless and free for the first 10,000 executions on Elastic Observability Serverless and Elastic Security Serverless. Executions are then as low as $0.025 per execution afterward.