# Elastic AI — AI Compare Ninja

> The AI layer of the Elastic Stack, with semantic search (ELSER), native vector search, RAG playground and AI Assistant; usage-based billing by compute, storage and inference tokens.

## Overview
- **Category:** Development & Code (Localization, search & APIs)
- **Status:** Active · checked - Active. Search AI is Elastic's AI layer on the Elastic Stack: semantic search, vector database, RAG playground, AI Assistant and agentic workflows for search, observability and security. Available as an open-source core (self-hosted), Elastic Cloud Hosted, or serverless.
- **Paid from:** from 0.09 $ / VCU-Stunde (Search-Compute) / Use-based, billed monthly (Elastic Cloud Serverless (Search))
- **Free access:** Free plan
- **Website:** https://www.elastic.co/search-ai
- **Canonical URL:** https://www.aicompare.ninja/en/tool/elastic-ai

## What it does
- Semantic search with ELSER and native vector search in the Elastic Stack
- Retrieval-augmented generation (RAG) playground and AI Assistant
- Agent Builder with 1,000 free executions per month
- Usage-based billing by compute, storage and inference tokens
- Workflows with a free allowance and a scalable Inference Service

## In brief
The AI layer of the Elastic Stack, with semantic search (ELSER), native vector search, RAG playground and AI Assistant; usage-based billing by compute, storage and inference tokens.

Serverless bills compute, storage, egress, and AI usage separately. Elastic lists $4.50 per million input tokens and $21 per million output tokens for Managed LLM; Workflows and Agent Builder each start with a monthly allowance. Elastic says it does not use customer data to train its own AI models or third-party models used through Managed LLM. Separately connected third-party models have their own data terms.


## Limits
Serverless bills compute, storage, egress, and AI usage separately. Elastic lists $4.50 per million input tokens and $21 per million output tokens for Managed LLM; Workflows and Agent Builder each start with a monthly allowance. Elastic says it does not use customer data to train its own AI models or third-party models used through Managed LLM. Separately connected third-party models have their own data terms.

## Pricing tiers
| Tier | Price | Period | What you get |
| --- | --- | --- | --- |
| Elastic Cloud Serverless (Search) | from 0.09 $ / VCU-Stunde (Search-Compute) | Use-based, billed monthly | Ingest compute from $0.14/VCU-hour, ML compute from $0.07/VCU-hour; Storage & retention from $0.047/GB/month; 50 GB storage for vector profiles included free; Egress from $0.05/GB; Elastic Managed LLM (AI Playground, Agent Builder, AI Assistant): $4.50 per 1M input tokens, $21 per 1M output tokens; Elastic Inference Service from $0.08 per 1M tokens (model-dependent); Workflows: 10,000 executions/month free, then from $0.0108/execution; Agent Builder: 1,000 executions/month free, then from $0.025/execution |

## Sources
All details regarding plans, limits, and features are based on official vendor documentation and the primary sources verified below.

- https://www.elastic.co/search-ai
- https://www.elastic.co/pricing/serverless-search
- https://www.elastic.co/trust/ai-data-faq

---
*Source: AI Compare Ninja (2026) - https://www.aicompare.ninja/en/tool/elastic-ai*