How to Intermediate · 4 min read

How to use Elasticsearch with Haystack

Q: How to use Elasticsearch with Haystack

Use Haystack with an ElasticsearchDocumentStore to index and retrieve documents, then connect it to a Retriever and Generator for AI-powered search. Install haystack-ai and elasticsearch, configure the document store, and run queries with the pipeline.

Quick answer

Use Haystack with an ElasticsearchDocumentStore to index and retrieve documents, then connect it to a Retriever and Generator for AI-powered search. Install haystack-ai and elasticsearch, configure the document store, and run queries with the pipeline.

PREREQUISITES

Python 3.8+
pip install haystack-ai>=2.0
Elasticsearch server running (7.x or 8.x)
OpenAI API key (free tier works)
pip install openai>=1.0

Setup

Install the required Python packages and ensure you have a running Elasticsearch instance. You can run Elasticsearch locally via Docker or use a managed service.

Install Haystack and Elasticsearch client:

bash

pip install haystack-ai elasticsearch openai

Step by step

This example shows how to create an ElasticsearchDocumentStore, index documents, and run a query using Haystack's Pipeline with OpenAI's GPT model for answer generation.

python

import os
from haystack import Pipeline
from haystack.document_stores import ElasticsearchDocumentStore
from haystack.nodes import OpenAIGenerator, BM25Retriever

# Configure Elasticsearch document store
document_store = ElasticsearchDocumentStore(
    host="localhost",
    port=9200,
    username="",
    password="",
    index="haystack-docs"
)

# Sample documents to index
docs = [
    {"content": "Elasticsearch is a distributed, RESTful search engine.", "meta": {"source": "wiki"}},
    {"content": "Haystack is a framework for building search systems.", "meta": {"source": "wiki"}}
]

# Write documents to Elasticsearch
document_store.write_documents(docs)

# Initialize retriever and generator
retriever = BM25Retriever(document_store=document_store)
generator = OpenAIGenerator(api_key=os.environ["OPENAI_API_KEY"], model="gpt-4o-mini")

# Build pipeline
pipeline = Pipeline()
pipeline.add_node(component=retriever, name="Retriever", inputs=["Query"])
pipeline.add_node(component=generator, name="Generator", inputs=["Retriever"])

# Run query
query = "What is Elasticsearch?"
result = pipeline.run(query=query, params={"Retriever": {"top_k": 3}})

print("Answer:", result["answers"][0].answer)

output

Answer: Elasticsearch is a distributed, RESTful search engine.

Common variations

You can use different retrievers like DPRRetriever for dense retrieval or switch to other generators such as OpenAIChatCompletion for chat-based models. Async pipelines are also supported in Haystack 2.x.

python

from haystack.nodes import OpenAIChatCompletion, DPRRetriever

# Example with DPRRetriever and chat model
retriever = DPRRetriever(document_store=document_store)
generator = OpenAIChatCompletion(api_key=os.environ["OPENAI_API_KEY"], model="gpt-4o")

pipeline = Pipeline()
pipeline.add_node(component=retriever, name="Retriever", inputs=["Query"])
pipeline.add_node(component=generator, name="Generator", inputs=["Retriever"])

result = pipeline.run(query="Explain Haystack framework", params={"Retriever": {"top_k": 5}})
print("Chat answer:", result["answers"][0].answer)

output

Chat answer: Haystack is an open-source framework designed to build powerful search systems that combine traditional search with AI models.

Troubleshooting

If you get connection errors, verify your Elasticsearch server is running and accessible at the configured host and port.
Ensure the Elasticsearch version is compatible with Haystack (7.x or 8.x).
If indexing fails, check document format and permissions.
For OpenAI errors, confirm your OPENAI_API_KEY environment variable is set correctly.

✅

Key Takeaways

Use ElasticsearchDocumentStore in Haystack to leverage Elasticsearch for document indexing and retrieval.
Combine retrievers like BM25Retriever with generators such as OpenAIGenerator for AI-powered search answers.
Ensure Elasticsearch is running and accessible before indexing or querying documents.
Haystack supports multiple retriever and generator variations for flexible search pipelines.
Set your OpenAI API key in os.environ["OPENAI_API_KEY"] to enable generation.

Verified 2026-04 · gpt-4o-mini, gpt-4o

Verify ↗