searxngRebrandZaclys/docs/admin/engines/search-indexer-engines.rst

Local Search Engines

further read

• Comparison to alternatives

Administrators might find themselves wanting to integrate locally running search engines. The following ones are supported for now:

• Elasticsearch
• Meilisearch
• Solr

Each search engine is powerful, capable of full-text search. All of the engines above are added to settings.yml just commented out, as you have to base_url for all them.

Please note that if you are not using HTTPS to access these engines, you have to enable HTTP requests by setting enable_http to True.

Futhermore, if you do not want to expose these engines on a public instance, you can still add them and limit the access by setting tokens as described in section private engines.

MeiliSearch

info

• meilisearch.py <searx/engines/meilisearch.py>
• MeiliSearch
• MeiliSearch Documentation
• Install MeiliSearch

MeiliSearch is aimed at individuals and small companies. It is designed for small-scale (less than 10 million documents) data collections. E.g. it is great for storing web pages you have visited and searching in the contents later.

The engine supports faceted search, so you can search in a subset of documents of the collection. Furthermore, you can search in MeiliSearch instances that require authentication by setting auth_token.

Here is a simple example to query a Meilisearch instance:

- name: meilisearch
  engine: meilisearch
  shortcut: mes
  base_url: http://localhost:7700
  index: my-index
  enable_http: true

Elasticsearch

info

• elasticsearch.py <searx/engines/elasticsearch.py>
• Elasticsearch
• Elasticsearch Guide
• Install Elasticsearch

Elasticsearch supports numerous ways to query the data it is storing. At the moment the engine supports the most popular search methods (query_type):

• match,
• simple_query_string,
• term and
• terms.

If none of the methods fit your use case, you can select custom query type and provide the JSON payload to submit to Elasticsearch in custom_query_json.

The following is an example configuration for an Elasticsearch instance with authentication configured to read from my-index index.

- name: elasticsearch
  shortcut: es
  engine: elasticsearch
  base_url: http://localhost:9200
  username: elastic
  password: changeme
  index: my-index
  query_type: match
  # custom_query_json: '{ ... }'
  enable_http: true

Solr

info

• solr.py <searx/engines/solr.py>
• Solr
• Solr Resources
• Install Solr

Solr is a popular search engine based on Lucene, just like Elasticsearch. But instead of searching in indices, you can search in collections.

This is an example configuration for searching in the collection my-collection and get the results in ascending order.

- name: solr
  engine: solr
  shortcut: slr
  base_url: http://localhost:8983
  collection: my-collection
  sort: asc
  enable_http: true

Acknowledgment

This development was sponsored by Search and Discovery Fund of NLnet Foundation.