How to Self-Host OpenSearch with Docker for Vector Search and Analytics

Getting Started with OpenSearch

Self-hosting OpenSearch is straightforward thanks to its official Docker image. Below is a complete guide to getting your own OpenSearch cluster running in minutes for AI-powered search and vector storage.

Prerequisites

Ensure Docker and Docker Compose are installed on your machine. OpenSearch requires at least 2 GB of RAM for basic operation and 4 GB recommended for production use. For vector search workloads handling embeddings from LLMs, allocate at least 8 GB.

Quick Start with Docker

docker pull opensearchproject/opensearch:latest
docker run -d --name opensearch-node   -p 9200:9200 -p 9600:9600   -e "discovery.type=single-node"   -e "OPENSEARCH_INITIAL_ADMIN_PASSWORD=YourSt...d1"   opensearchproject/opensearch:latest

Once running, verify by curling the API:

curl -k -u admin:YourStrong@Password1 https://localhost:9200

Docker Compose for Production

For a more robust setup with OpenSearch Dashboards, use docker-compose:

services:
  opensearch:
    image: opensearchproject/opensearch:latest
    restart: unless-stopped
    ports:
      - 9200:9200
    volumes:
      - ./data/opensearch:/data
    environment:
      - discovery.type=single-node
      - OPENSEARCH_INITIAL_ADMIN_PASSWORD=YourSt...

  opensearch-dashboards:
    image: opensearchproject/opensearch-dashboards:latest
    ports:
      - 5601:5601
    environment:
      - OPENSEARCH_HOSTS=https://opensearch:9200

Vector Search Quick Start

Once OpenSearch is running, create an index with a vector field for your embeddings:

PUT /my-rag-index
{
  "settings": { "index": { "knn": true } },
  "mappings": {
    "properties": {
      "content": { "type": "text" },
      "embedding": { "type": "knn_vector", "dimension": 768 }
    }
  }
}

Why Self-Host OpenSearch?

Running your own OpenSearch gives you full control over your data, no licensing costs, unlimited storage, and the ability to tune performance for your specific AI workloads. It integrates with every major RAG framework and supports hybrid search for the best retrieval quality.

Conclusion

OpenSearch is production-ready, feature-rich, and completely free. With Docker, you can have a vector search engine powering your AI applications in under 5 minutes. Visit the Run This Ai tool page for the full docker-compose setup.