Milvus Embedding Store

Milvus is a scalable and high-performance vector database optimized for AI and semantic search use cases. This guide explains how to use Milvus as an embedding store in Quarkus LangChain4j for RAG applications.

Overview

The quarkus-langchain4j-milvus extension enables seamless integration with a Milvus instance for storing and retrieving embedded documents using vector similarity.

Milvus supports approximate nearest neighbor (ANN) search with various indexing strategies and similarity metrics.

Dependency

To enable Milvus integration in your Quarkus project, add the following dependency:

<dependency>
  <groupId>io.quarkiverse.langchain4j</groupId>
  <artifactId>quarkus-langchain4j-milvus</artifactId>
  <version>1.10.0</version>
</dependency>

Even better, if you use the Quarkus platform BOM (default for projects generated), add the Quarkus Langchain4J BOM and all dependency versions will align:

    <dependencyManagement>
        <dependencies>
            <dependency>
                <groupId>${quarkus.platform.group-id}</groupId>
                <artifactId>${quarkus.platform.artifact-id}</artifactId>
                <version>${quarkus.platform.version}</version>
                <type>pom</type>
                <scope>import</scope>
            </dependency>
            <dependency>
                <groupId>${quarkus.platform.group-id}</groupId>
                <artifactId>quarkus-langchain4j-bom</artifactId> (1)
                <version>${quarkus.platform.version}</version> (2)
                <type>pom</type>
                <scope>import</scope>
            </dependency>
        </dependencies>
    </dependencyManagement>

    <dependencies>
      <dependency>
        <groupId>io.quarkiverse.langchain4j</groupId>
        <artifactId>quarkus-langchain4j-milvus</artifactId>
        (3)
      </dependency>
    </dependencies>
1 In your dependencyManagement section, add the quarkus-langchain4j-bom
2 Inherit the version from your platform version
3 Voilà, no need for version alignment anymore

Dev Services Support

This extension includes Dev Services support. In dev and test mode, a containerized Milvus instance is started automatically.

To configure Dev Services for Milvus, set the vector dimension according to your embedding model:

quarkus.langchain4j.milvus.dimension=384
For example, AllMiniLmL6V2QuantizedEmbeddingModel → 384; OpenAI text-embedding-ada-002 → 1536.

Connecting to an External Milvus Instance

If you prefer to connect to a remote Milvus instance, disable Dev Services and provide the host and port:

quarkus.langchain4j.milvus.host=localhost
quarkus.langchain4j.milvus.port=19530
quarkus.langchain4j.milvus.dimension=384

When a host is defined, Dev Services will not start a container.

Usage Example

Once the extension is configured, you can use Milvus as an embedding store in your ingestion logic:

package io.quarkiverse.langchain4j.samples;

import static dev.langchain4j.data.document.splitter.DocumentSplitters.recursive;

import java.util.List;

import jakarta.enterprise.context.ApplicationScoped;
import jakarta.inject.Inject;

import dev.langchain4j.data.document.Document;
import dev.langchain4j.model.embedding.EmbeddingModel;
import dev.langchain4j.store.embedding.EmbeddingStoreIngestor;
import dev.langchain4j.store.embedding.milvus.MilvusEmbeddingStore;

@ApplicationScoped
public class IngestorExampleWithMilvus {

    /**
     * The embedding store (the database).
     * The bean is provided by the quarkus-langchain4j-milvus extension.
     */
    @Inject
    MilvusEmbeddingStore store;

    /**
     * The embedding model (how is computed the vector of a document).
     * The bean is provided by the LLM (like openai) extension.
     */
    @Inject
    EmbeddingModel embeddingModel;

    public void ingest(List<Document> documents) {
        EmbeddingStoreIngestor ingestor = EmbeddingStoreIngestor.builder()
                .embeddingStore(store)
                .embeddingModel(embeddingModel)
                .documentSplitter(recursive(500, 0))
                .build();
        // Warning - this can take a long time...
        ingestor.ingest(documents);
    }
}

This example shows how to store vectorized content and make it retrievable via similarity search.

Configuration Options

Customize the behavior of the extension using the following configuration options:

Configuration property fixed at build time - All other configuration properties are overridable at runtime

Configuration property

Type

Default

quarkus.langchain4j.milvus.default-store-enabled

Whether the default (unnamed) Milvus embedding store should be enabled. Set to false when you only want to use named stores.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS_DEFAULT_STORE_ENABLED

boolean

true

quarkus.langchain4j.milvus.devservices.enabled

Whether Dev Services for Milvus are enabled or not.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS_DEVSERVICES_ENABLED

boolean

true

quarkus.langchain4j.milvus.devservices.milvus-image-name

Container image for Milvus.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS_DEVSERVICES_MILVUS_IMAGE_NAME

string

docker.io/milvusdb/milvus:v2.3.16

quarkus.langchain4j.milvus.devservices.port

Optional fixed port the Milvus dev service will listen to. If not defined, the port will be chosen randomly.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS_DEVSERVICES_PORT

int

quarkus.langchain4j.milvus.devservices.shared

Indicates if the Dev Service containers managed by Quarkus for Milvus are shared.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS_DEVSERVICES_SHARED

boolean

true

quarkus.langchain4j.milvus.devservices.service-name

Service label to apply to created Dev Services containers.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS_DEVSERVICES_SERVICE_NAME

string

milvus

quarkus.langchain4j.milvus.host

The URL of the Milvus server.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS_HOST

string

localhost

quarkus.langchain4j.milvus.port

The port of the Milvus server.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS_PORT

int

19530

quarkus.langchain4j.milvus.token

The authentication token for the Milvus server.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS_TOKEN

string

quarkus.langchain4j.milvus.username

The username for the Milvus server.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS_USERNAME

string

quarkus.langchain4j.milvus.password

The password for the Milvus server.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS_PASSWORD

string

quarkus.langchain4j.milvus.timeout

The timeout duration for the Milvus client. If not specified, 5 seconds will be used.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS_TIMEOUT

Duration 

quarkus.langchain4j.milvus.db-name

Name of the database.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS_DB_NAME

string

default

quarkus.langchain4j.milvus.create-collection

Create the collection if it does not exist yet.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS_CREATE_COLLECTION

boolean

true

quarkus.langchain4j.milvus.collection-name

Name of the collection.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS_COLLECTION_NAME

string

embeddings

quarkus.langchain4j.milvus.dimension

Dimension of the vectors. Only applicable when the collection yet has to be created.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS_DIMENSION

int

quarkus.langchain4j.milvus.primary-field

Name of the field that contains the ID of the vector.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS_PRIMARY_FIELD

string

id

quarkus.langchain4j.milvus.text-field

Name of the field that contains the text from which the vector was calculated.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS_TEXT_FIELD

string

text

quarkus.langchain4j.milvus.metadata-field

Name of the field that contains JSON metadata associated with the text.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS_METADATA_FIELD

string

metadata

quarkus.langchain4j.milvus.vector-field

Name of the field to store the vector in.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS_VECTOR_FIELD

string

vector

quarkus.langchain4j.milvus.description

Description of the collection.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS_DESCRIPTION

string

quarkus.langchain4j.milvus.index-type

The index type to use for the collection.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS_INDEX_TYPE

none, flat, ivf-flat, ivf-sq8, ivf-pq, hnsw, hnsw-sq, hnsw-pq, hnsw-prq, diskann, autoindex, scann, gpu-ivf-flat, gpu-ivf-pq, gpu-brute-force, gpu-cagra, bin-flat, bin-ivf-flat, trie, stl-sort, inverted, bitmap, sparse-inverted-index, sparse-wand

flat

quarkus.langchain4j.milvus.metric-type

The metric type to use for searching.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS_METRIC_TYPE

none, l2, ip, cosine, hamming, jaccard

cosine

quarkus.langchain4j.milvus.consistency-level

The consistency level.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS_CONSISTENCY_LEVEL

strong, session, bounded, eventually

eventually

Named store configurations

Type

Default

quarkus.langchain4j.milvus."store-name".host

The host of the Milvus server to use for this named store. If set to <default>, the default Milvus server will be used.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS__STORE_NAME__HOST

string

quarkus.langchain4j.milvus."store-name".host

The URL of the Milvus server.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS__STORE_NAME__HOST

string

localhost

quarkus.langchain4j.milvus."store-name".port

The port of the Milvus server.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS__STORE_NAME__PORT

int

19530

quarkus.langchain4j.milvus."store-name".token

The authentication token for the Milvus server.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS__STORE_NAME__TOKEN

string

quarkus.langchain4j.milvus."store-name".username

The username for the Milvus server.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS__STORE_NAME__USERNAME

string

quarkus.langchain4j.milvus."store-name".password

The password for the Milvus server.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS__STORE_NAME__PASSWORD

string

quarkus.langchain4j.milvus."store-name".timeout

The timeout duration for the Milvus client. If not specified, 5 seconds will be used.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS__STORE_NAME__TIMEOUT

Duration 

quarkus.langchain4j.milvus."store-name".db-name

Name of the database.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS__STORE_NAME__DB_NAME

string

default

quarkus.langchain4j.milvus."store-name".create-collection

Create the collection if it does not exist yet.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS__STORE_NAME__CREATE_COLLECTION

boolean

true

quarkus.langchain4j.milvus."store-name".collection-name

Name of the collection.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS__STORE_NAME__COLLECTION_NAME

string

embeddings

quarkus.langchain4j.milvus."store-name".dimension

Dimension of the vectors. Only applicable when the collection yet has to be created.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS__STORE_NAME__DIMENSION

int

quarkus.langchain4j.milvus."store-name".primary-field

Name of the field that contains the ID of the vector.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS__STORE_NAME__PRIMARY_FIELD

string

id

quarkus.langchain4j.milvus."store-name".text-field

Name of the field that contains the text from which the vector was calculated.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS__STORE_NAME__TEXT_FIELD

string

text

quarkus.langchain4j.milvus."store-name".metadata-field

Name of the field that contains JSON metadata associated with the text.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS__STORE_NAME__METADATA_FIELD

string

metadata

quarkus.langchain4j.milvus."store-name".vector-field

Name of the field to store the vector in.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS__STORE_NAME__VECTOR_FIELD

string

vector

quarkus.langchain4j.milvus."store-name".description

Description of the collection.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS__STORE_NAME__DESCRIPTION

string

quarkus.langchain4j.milvus."store-name".index-type

The index type to use for the collection.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS__STORE_NAME__INDEX_TYPE

none, flat, ivf-flat, ivf-sq8, ivf-pq, hnsw, hnsw-sq, hnsw-pq, hnsw-prq, diskann, autoindex, scann, gpu-ivf-flat, gpu-ivf-pq, gpu-brute-force, gpu-cagra, bin-flat, bin-ivf-flat, trie, stl-sort, inverted, bitmap, sparse-inverted-index, sparse-wand

flat

quarkus.langchain4j.milvus."store-name".metric-type

The metric type to use for searching.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS__STORE_NAME__METRIC_TYPE

none, l2, ip, cosine, hamming, jaccard

cosine

quarkus.langchain4j.milvus."store-name".consistency-level

The consistency level.

Environment variable: QUARKUS_LANGCHAIN4J_MILVUS__STORE_NAME__CONSISTENCY_LEVEL

strong, session, bounded, eventually

eventually
About the Duration format

To write duration values, use the standard java.time.Duration format. See the Duration#parse() Java API documentation for more information.

You can also use a simplified format, starting with a number:

  • If the value is only a number, it represents time in seconds.

  • If the value is a number followed by ms, it represents time in milliseconds.

In other cases, the simplified format is translated to the java.time.Duration format for parsing:

  • If the value is a number followed by h, m, or s, it is prefixed with PT.

  • If the value is a number followed by d, it is prefixed with P.

How It Works

Milvus stores each embedded text segment along with its metadata and vector in a collection. The vector similarity search uses one of several distance metrics:

  • COSINE

  • L2 (Euclidean)

  • IP (Inner Product / dot product)

The extension automatically initializes the collection and schema if it does not already exist. It uses gRPC-based access for high performance and low latency.

Named Stores

You can configure multiple named Milvus stores, each pointing to a different Milvus instance or collection. This is useful when your application needs to manage embeddings for different domains or tenants separately.

To configure a named store:

quarkus.langchain4j.milvus.products.host=<default>
quarkus.langchain4j.milvus.products.dimension=1536
quarkus.langchain4j.milvus.products.collection-name=product_embeddings

Setting host to <default> means the named store shares the same Milvus instance as the default store. To point to a different Milvus instance, specify its host and port:

quarkus.langchain4j.milvus.products.host=another-milvus.internal
quarkus.langchain4j.milvus.products.port=19530
quarkus.langchain4j.milvus.products.dimension=1536

To inject a named store, use the @EmbeddingStoreName qualifier:

@Inject
@EmbeddingStoreName("products")
MilvusEmbeddingStore productsStore;

The default store and named stores can coexist. If you only need named stores, disable the default store:

quarkus.langchain4j.milvus.default-store-enabled=false

Summary

To use Milvus with Quarkus LangChain4j:

  1. Add the required extension

  2. Set the embedding dimension to match your model

  3. Use Dev Services or connect to an external Milvus instance

  4. Ingest documents and perform similarity search using MilvusEmbeddingStore