OpenCrawling
← Back to Blog Articles
Feature Release September 3, 2026 • 6 min read

Solr 10 Dense Vector Search & HNSW Quantization: Announcing the Apache Solr Output Connector

We are thrilled to announce the official release of the Apache Solr Output Connector (oc-solr-output-connector)! OpenCrawling now natively streams enterprise document repositories into Apache Solr 10.x and 9.x—one of the world's most trusted open-source search platforms—delivering high-throughput vector chunk indexing, HNSW ANN k-nearest neighbor search, Solr 10 vector quantization, dynamic REST schema auto-provisioning, and zero-trust ACL security filtering.

01. Why Apache Solr 10 for Enterprise Vector Search?

Apache Solr 10 introduces major advances in dense vector search. By natively supporting solr.DenseVectorField with HNSW indexing, Solr combines world-class full-text search with modern vector similarity scoring.

💡

SolrJ 10 JDK HTTP Client: Solr 10 replaces legacy Jetty client dependencies with Java's standard java.net.http.HttpClient (HttpJdkSolrClient / CloudSolrClient), eliminating classpath library conflicts and ensuring high-performance asynchronous HTTP connections.

02. End-to-End Pipeline Architecture

Decoupled Ingestion Pipeline → Apache Solr Collection
Repository Crawlers (Filesystem, ACS, Iceberg, Flowable)
→
Kafka (opencrawling-embedded)
→
SolrStoreWriterConsumer (oc-solr-output-connector)
→
Solr 10 solr.DenseVectorField ('enterprise_kb')

03. Automatic REST Schema Auto-Provisioning

Upon initialization, SolrCollectionInitializer checks whether the target Solr collection contains the required knn_vector field type and schema fields. If missing, it automatically posts the field type definition to Solr's Schema API: 

{
  "add-field-type": {
    "name": "knn_vector",
    "class": "solr.DenseVectorField",
    "vectorDimension": 1024,
    "similarityFunction": "cosine",
    "knnAlgorithm": "hnsw",
    "hnswMaxConnections": 16,
    "hnswBeamWidth": 100,
    "vectorEncoding": "FLOAT32"
  },
  "add-field": [
    { "name": "id", "type": "string", "stored": true, "indexed": true },
    { "name": "text", "type": "text_general", "stored": true, "indexed": true },
    { "name": "embeddings", "type": "knn_vector", "stored": true, "indexed": true },
    { "name": "security_allowed_read", "type": "string", "stored": true, "indexed": true, "multiValued": true },
    { "name": "security_denied_read", "type": "string", "stored": true, "indexed": true, "multiValued": true }
  ]
}

04. Configuration Properties

All properties are prefixed with spring.opencrawling.output.solr.* and support environment variable overrides:

Property Default Description
spring.opencrawling.output.solr.mode standalone Solr deployment mode: standalone or cloud
spring.opencrawling.output.solr.url http://localhost:8983/solr Base HTTP URL of standalone Solr instance
spring.opencrawling.output.solr.collection enterprise_kb Target Solr collection / core name
spring.opencrawling.output.solr.dimensions 1024 Vector embedding dimension size
spring.opencrawling.output.solr.vector-encoding FLOAT32 Solr 10 vector encoding: FLOAT32 or BYTE
spring.opencrawling.output.solr.quantization none Solr 10 quantization: none, scalar, or binary

05. Dense Vector Search Queries with Solr 10 KNN Parser

Executing hybrid and vector similarity searches against Apache Solr is straightforward using Solr's {!knn} query parser: 

// Full-Text Search
curl "http://localhost:8983/solr/enterprise_kb/select?q=text:OpenCrawling&fl=id,text,uri,score"

// Dense Vector KNN Search (Solr 10)
curl --data-urlencode "q={!knn f=embeddings topK=5}[0.0123,-0.0456,0.0891,...]" \
     --data-urlencode "fl=id,text,uri,score" \
     "http://localhost:8983/solr/enterprise_kb/select"

06. Getting Started & Integration Testing

You can spin up Apache Solr and run the complete decoupled integration pipeline with one command: 

./scripts/test-solr-decoupled.sh

Ready to Index Enterprise Documents into Apache Solr 10?

Explore the full connector source code, build custom RAG pipelines, or read our wiki guide.

View Module Code Read Solr Wiki Guide