Hybrid Search & AI Search for Solr — Guides

Using NLP Models

Using NLP Models in Your Solr `schema_extra_types.xml`

Leverage the power of Natural Language Processing (NLP) right inside Solr!
With built-in support for OpenNLP models, you can add advanced tokenization, part-of-speech tagging, named entity recognition, and much more—no PhD required.

Why Use NLP Models in Solr?

Integrating NLP in your schema allows you to:

Extract nouns, verbs, or any part-of-speech you fancy.
Perform more relevant searches by filtering, stemming, and synonymizing.
Create blazing-fast autocomplete and suggestion features via EdgeNGrams.
Support multi-language, linguistically smart queries.

In short: your Solr becomes smarter and your users get better search results.

Example: Dutch Edge NGram Nouns Field

Here’s a typical fieldType in your schema_extra_types.xml using OpenNLP:

<fieldType name="text_edge_nouns_nl" class="solr.TextField" positionIncrementGap="100">
  <analyzer type="index">
    <tokenizer class="solr.OpenNLPTokenizerFactory" sentenceModel="/opt/nlp/nl-sent.bin" tokenizerModel="/opt/nlp/nl-token.bin"/>
    <filter class="solr.OpenNLPPOSFilterFactory" posTaggerModel="/opt/nlp/nl-pos-maxent.bin"/>
    <filter class="solr.TypeTokenFilterFactory" types="pos_edge_nouns_nl.txt" useWhitelist="true"/>
    <filter class="solr.LowerCaseFilterFactory"/>
    <filter class="solr.RemoveDuplicatesTokenFilterFactory"/>
    <filter class="solr.EdgeNGramFilterFactory" minGramSize="2" maxGramSize="25"/>
  </analyzer>
  <analyzer type="query">
    <tokenizer class="solr.OpenNLPTokenizerFactory" sentenceModel="/opt/nlp/nl-sent.bin" tokenizerModel="/opt/nlp/nl-token.bin"/>
    <filter class="solr.OpenNLPPOSFilterFactory" posTaggerModel="/opt/nlp/nl-pos-maxent.bin"/>
    <filter class="solr.TypeTokenFilterFactory" types="pos_edge_nouns_nl.txt" useWhitelist="true"/>
    <filter class="solr.LowerCaseFilterFactory"/>
    <filter class="solr.SynonymGraphFilterFactory" synonyms="synonyms_edge_nouns_nl.txt"/>
    <filter class="solr.RemoveDuplicatesTokenFilterFactory"/>
  </analyzer>
</fieldType>

Important Details

Model Paths:
Always reference the full absolute path for NLP model files. For example:
```
sentenceModel="/opt/nlp/nl-sent.bin"
tokenizerModel="/opt/nlp/nl-token.bin"
posTaggerModel="/opt/nlp/nl-pos-maxent.bin"
```
This ensures Solr always finds your precious language models—no “file not found” drama!
Type Token Filtering:
The TypeTokenFilterFactory with useWhitelist="true" will only keep tokens matching the allowed parts of speech (like nouns, verbs, etc.), as defined in pos_edge_nouns_nl.txt. This keeps your index tight and focused.
Synonym Graphs:
Add SynonymGraphFilterFactory to enable query-side expansion. This is great for handling multiple word forms, synonyms, and local lingo.

‍Best Practices & Gotchas

Keep your NLP model files up to date and tested for your language version!
If using multiple languages, make sure you have the right models for each language. (No, Dutch models won’t help with Klingon. Yet.)
EdgeNGram and NGram fields are fantastic for autocomplete—but don’t overdo it, as they can bloat your index if not tuned.
Use RemoveDuplicatesTokenFilterFactory to keep things clean and efficient.

Not Just for Dutch!

You can set up similar analyzers for English, undefined language, or anything you like. For example:

<fieldType name="text_nouns_en" class="solr.TextField" positionIncrementGap="100">
  <analyzer type="index">
    <tokenizer class="solr.OpenNLPTokenizerFactory" sentenceModel="/opt/nlp/en-sent.bin" tokenizerModel="/opt/nlp/en-token.bin"/>
    <filter class="solr.OpenNLPPOSFilterFactory" posTaggerModel="/opt/nlp/en-pos-maxent.bin"/>
    <filter class="solr.TypeTokenFilterFactory" types="pos_nouns_en.txt" useWhitelist="true"/>
    <filter class="solr.LowerCaseFilterFactory"/>
    <filter class="solr.RemoveDuplicatesTokenFilterFactory"/>
  </analyzer>
  <analyzer type="query">
    <tokenizer class="solr.OpenNLPTokenizerFactory" sentenceModel="/opt/nlp/en-sent.bin" tokenizerModel="/opt/nlp/en-token.bin"/>
    <filter class="solr.OpenNLPPOSFilterFactory" posTaggerModel="/opt/nlp/en-pos-maxent.bin"/>
    <filter class="solr.TypeTokenFilterFactory" types="pos_nouns_en.txt" useWhitelist="true"/>
    <filter class="solr.LowerCaseFilterFactory"/>
    <filter class="solr.SynonymGraphFilterFactory" synonyms="synonyms_nouns_en.txt"/>
    <filter class="solr.RemoveDuplicatesTokenFilterFactory"/>
  </analyzer>
</fieldType>

Keep It Organized

Store all model files in a single, logical directory (like /opt/nlp/), and keep a README so you know what’s what.
Protect those models! They’re your “brains” for language tasks.

Wrap-up

Using NLP models in your Solr analyzers will supercharge your search, make autocomplete smarter, and help users find what they’re actually looking for (even if they type like my cat walks on a keyboard).

Need more examples?
Check out the Solr Reference Guide - OpenNLP Integration or the OpenNLP NER guide for Opensolr.

Happy indexing, and may your tokens always be well-typed!

Read Full Answer

How to use OpenNLP (NER) with Opensolr

OpenNLP + Opensolr Integration Guide

Heads up!
Before you dive into using NLP models with your Opensolr index, please contact us to request the NLP models to be installed for your Opensolr index.
We'll reply with the correct path to use for the .bin files in your schema.xml or solrconfig.xml.

What’s this all about?

This is your step-by-step guide to using OpenNLP models with Opensolr. In this walkthrough, we’ll cover Named Entity Recognition (NER) using default OpenNLP models, so you can start extracting valuable information (like people, places, and organizations) directly from your indexed data.

Note:
Currently, these models are enabled by default only in the Germany, Solr Version 9 environment. So, if you want an easy life, create your index there!
We’re happy to set up the models in any region (or even your dedicated Opensolr infrastructure for corporate accounts) if you reach out via our contact form.

The default OpenNLP models are on the official OpenNLP website.

Step-by-Step: Enable NLP Entity Extraction

Create your Opensolr Index
- Use this guide to create your Opensolr index (Solr 7, 8 or 9; Solr versions below 8 are retired on 01/01/2027).
- Pro Tip: Creating your index in the Germany Solr 9 Web Crawler Environment skips most of the manual steps below.
Edit Your schema.xml
- Go to the Opensolr Control Panel.
- Click your Index Name → Configuration tab → select schema.xml to edit.
- Add these snippets:
  
  Dynamic Field (for storing entities):

<dynamicField name="*_s" type="string" multiValued="true" indexed="true" stored="true" termVectors="true" termPositions="true" termOffsets="true" storeOffsetsWithPositions="true" />

  **NLP Tokenizer fieldType:**

<fieldType name="text_nlp" class="solr.TextField" positionIncrementGap="100">
    <analyzer>
        <charFilter class="solr.MappingCharFilterFactory" mapping="mapping-ISOLatin1Accent.txt"/>
        <tokenizer class="solr.OpenNLPTokenizerFactory"
            sentenceModel="en-sent.bin"
            tokenizerModel="en-token.bin"/>
         <filter class="solr.OpenNLPPOSFilterFactory" posTaggerModel="en-pos-maxent.bin"/>
         <filter class="solr.OpenNLPChunkerFilterFactory" chunkerModel="en-chunker.bin"/>
         <filter class="solr.TypeAsPayloadFilterFactory"/>
     </analyzer>
 </fieldType>

- **Important:** Don’t use the `text_nlp` type for your dynamic fields! It’s only for the update processor.

Save, then Edit Your solrconfig.xml
- Add the following updateRequestProcessorChain (and corresponding requestHandler):

<requestHandler name="/update" class="solr.UpdateRequestHandler" >
    <lst name="defaults">
        <str name="update.chain">nlp</str>
    </lst>
</requestHandler>
<updateRequestProcessorChain name="nlp">
    <!-- Extract English People Names -->
    <processor class="solr.OpenNLPExtractNamedEntitiesUpdateProcessorFactory">
        <str name="modelFile">en-ner-person.bin</str>
        <str name="analyzerFieldType">text_nlp</str>
        <arr name="source">
            <str>title</str>
            <str>description</str>
        </arr>
        <str name="dest">people_s</str>
    </processor>
    <!-- Extract Spanish People Names -->
    <processor class="solr.OpenNLPExtractNamedEntitiesUpdateProcessorFactory">
        <str name="modelFile">es-ner-person.bin</str>
        <str name="analyzerFieldType">text_nlp</str>
        <arr name="source">
            <str>title</str>
            <str>description</str>
        </arr>
        <str name="dest">people_s</str>
    </processor>
    <!-- Extract Locations -->
    <processor class="solr.OpenNLPExtractNamedEntitiesUpdateProcessorFactory">
        <str name="modelFile">en-ner-location.bin</str>
        <str name="analyzerFieldType">text_nlp</str>
        <arr name="source">
            <str>title</str>
            <str>description</str>
        </arr>
        <str name="dest">location_s</str>
    </processor>
    <!-- Extract Organizations -->
    <processor class="solr.OpenNLPExtractNamedEntitiesUpdateProcessorFactory">
        <str name="modelFile">en-ner-organization.bin</str>
        <str name="analyzerFieldType">text_nlp</str>
        <arr name="source">
            <str>title</str>
            <str>description</str>
        </arr>
        <str name="dest">organization_s</str>
    </processor>
    <!-- Language Detection -->
    <processor class="org.apache.solr.update.processor.OpenNLPLangDetectUpdateProcessorFactory">
        <str name="langid.fl">title,text,description</str>
        <str name="langid.langField">language_s</str>
        <str name="langid.model">langdetect-183.bin</str>
    </processor>
    <!-- Remove duplicate extracted entities -->
    <processor class="solr.UniqFieldsUpdateProcessorFactory">
        <str name="fieldRegex">.*_s</str>
    </processor>
    <processor class="solr.RunUpdateProcessorFactory" />
</updateRequestProcessorChain>

Populate Test Data (for the impatient!)
- If you’re using the Germany Solr 9 Web Crawler, you can crawl your site and extract all the juicy entities automatically.
- Or, insert a sample doc via Solr Admin:
Sample JSON:

{
    "id": "1",
    "title": "Jack Sparrow was a pirate. Many feared him. He used to live in downtown Las Vegas.",
    "description": "Jack Sparrow and Janette Sparrowa, are now on their way to Monte Carlo for the summer vacation, after working hard for Microsoft, creating the new and exciting Windows 11 which everyone now loves. :)",
    "text": "The Apache OpenNLP project is developed by volunteers and is always looking for new contributors to work on all parts of the project. Every contribution is welcome and needed to make it better. A contribution can be anything from a small documentation typo fix to a new component.Learn more about how you can get involved."
}

See the Magic!
- Visit the query tab to see extracted entities in action!

Related: Using NLP Models, a worked OpenNLP fieldType for part of speech tagging and autocomplete.

Read Full Answer

Hybrid Search Live Demo — Try It Yourself

Did you know? You can now use the {!hybrid} parser — a custom Opensolr Solr query parser that gives you better hybrid search: correct keyword + vector fusion, real search modes, and natural-language, cross-lingual queries out of the box. Learn more →

Live Search Demo

Hybrid Search in Action

Keyword matching + semantic understanding, working together. Click any query below to run it live — no sign-up needed.

How It Works

Keyword Layer

Finds docs with matching terms — fast, precise, field-weighted

Semantic Layer

Finds docs by meaning — works even with zero lexical overlap

Score Fusion

Our own private {!hybrid} parser blends keyword and meaning scores per document, tuned with one alpha knob and four modes

Zero Config

Automatic on all vector-enabled indexes — no query changes needed

Four Indexes, Four Domains

Each index below is a real, live Opensolr instance. Every query was chosen to demonstrate semantic understanding — words in the query often do not appear in the top-ranked documents.

Opensolr Docs

Technical knowledge base

Fluke

Test & measurement equipment

Office Supplies

B2B catalog (DE)

Starbucks at Home

Specialty coffee products

Opensolr Knowledge Base

150k+ documentation pages · English · Tech/SaaS domain

Query

"pricing plans and costs"

Surfaces the pricing page and plan comparison docs — semantic layer understands "costs" maps to "pricing" context.

Try it live →

Query

"crawl and index my website automatically"

Reaches Web Crawler feature docs — semantic understands "crawl automatically" as an intent for automated ingestion pipelines.

Try it live →

Query

"my search results are not relevant enough"

Finds relevance tuning and field weight documentation — the semantic layer catches the intent behind a user complaint phrasing.

Try it live →

Fluke Test & Measurement

150k+ product & support pages · English (locale=en_us) · Industrial domain

Semantic challenge: These queries are written in natural conversational language — they contain no part numbers, product names, or catalog terms. Pure semantic understanding required.

Query

"how do I check current without touching wires"

Finds clamp meters and non-contact current probes — the product category name never appears in the query, yet vector search retrieves exactly the right tools.

Try it live →

Query

"see heat through walls"

Returns thermal imaging cameras — the query contains zero product terminology. "See heat through walls" is pure human intent, correctly mapped to infrared thermography.

Try it live →

Query

"electrical safety inspection building"

Surfaces insulation resistance testers, installation testers, and safety guides — semantic layer understands the professional context of a building electrical inspector.

Try it live →

Office Supplies Catalog

German-language catalog · English queries · B2B office supplies

Cross-language semantic: Every product is listed in German and every query below is in English. They share no words, and multilingual E5 embeddings still match them.

Query (English)

"paper shredder"

Returns HSM shredstar paper shredders listed as Aktenvernichter: the English word never appears in the catalog.

Try it live →

Query (English)

"ergonomic office chair"

Returns ergonomic seating stools listed in German as Sitzhocker and Sitz-Steh-Hocker.

Try it live →

Query (English)

"something to hold loose papers together"

A description, not a product name: returns paper holders, clipboards and paper clip dispensers.

Try it live →

Starbucks at Home

400+ specialty coffee & beverage products · English · CPG domain

Dense catalog, small index: With only ~400 products, this demo shows how semantic search extracts relevant results even when keyword overlap is minimal or zero.

Query

"cozy winter hot chocolate"

Returns hot cocoa mixes and warming winter drinks — vector embeddings understand "cozy winter" as a seasonal mood context that maps perfectly to comfort beverage products.

Try it live →

Query

"morning pick me up caffeine boost"

Surfaces dark roast espresso pods and high-caffeine blends — semantic model understands "pick me up" as an energy/alertness intent, not a product descriptor.

Try it live →

Query

"sustainable farming environmental impact"

Finds ethically sourced and Rainforest Alliance certified coffees — no single query word appears in the product titles, yet vector search pinpoints environmentally responsible products.

Try it live →

Ready to add hybrid search to your index?

One flip. Instant semantic search.

Create your index in a vector region, or install the Opensolr vector schema on one with apply_schema, and hybrid search activates automatically. No query changes: the documents you send are embedded for you.

Browse Documentation Hybrid Search Deep Dive

Read Full Answer

Hybrid Search

Using NLP Models in Your Solr schema_extra_types.xml

Why Use NLP Models in Solr?

Example: Dutch Edge NGram Nouns Field

Important Details

‍Best Practices & Gotchas

Not Just for Dutch!

Keep It Organized

Wrap-up

OpenNLP + Opensolr Integration Guide

What’s this all about?

Step-by-Step: Enable NLP Entity Extraction

Hybrid Search in Action

How It Works

Four Indexes, Four Domains

One flip. Instant semantic search.

Using NLP Models in Your Solr `schema_extra_types.xml`