fess is a free, open source data extraction & web scraping project written in Java and released under Apache-2.0. It has 1,142 GitHub stars, 177 forks and 15 open issues, and was last pushed 18 hours ago. On this registry it ranks #113 of 150 tracked projects in Data Extraction & Web Scraping, with 5 head-to-head comparisons available. It gained 4 stars over the last 30 tracked days.

What is fess?

Fess is an open-source, self-hosted enterprise and site search server built on OpenSearch, intended for teams and site owners who need to crawl, index, and search their own web, file, database, and cloud content without sending it to a hosted search provider.

What it is

Fess is a Java-based enterprise search server, licensed under Apache-2.0, that can be installed and run on any platform with a Java runtime. It is built on OpenSearch, but prior OpenSearch knowledge is not required, because Fess is configured through a browser-based administration UI rather than through index-level tooling. A built-in crawler collects documents from web sites, file systems, and data stores such as databases and CSV files, and the resulting index is queried through a search UI and a REST API. The project sits in the same ecosystem as OpenSearch, Elasticsearch, and Lucene, and publishes Docker images as well as DEB, RPM, and ZIP packages.

The concrete problem it solves is standing up search over content that an organization already owns and already stores in many places. Instead of assembling a crawler, an index, an administration interface, and a permission model separately, Fess ships them as one server with configurable crawling targets. For public-facing search, the related Fess Site Search project provides a free alternative to Google Site Search that can be embedded in a website through the FSS JS Generator. The repository also carries a MIGRATION.md document for moving from another search provider.

Key capabilities

  • Full-text search with faceting, sorting, and search suggestions.
  • Crawlers for web sites, file systems, and data stores, including databases, cloud storage, and SaaS.
  • Data store connectors published as separate modules: Confluence/Jira, Box, CSV, Database, Dropbox, Elasticsearch, Git, Gitbucket, and G Suite.
  • Document format support including Microsoft Office, PDF, and ZIP archives.
  • Browser-based administration UI plus a REST API.
  • Role- and permission-based filtering of search results.
  • Single sign-on with LDAP, OpenID Connect, SAML, SPNEGO, and Microsoft Entra ID.
  • Multilingual user interface and text analysis across 20+ languages.
  • Extensibility through data store, ingest, script, and theme plugins.

Who uses it and how

  • Organizations that need internal search across Confluence/Jira, Box, Dropbox, Git, and G Suite content, using the published data store connectors to register each source.
  • Public site owners who want embedded site search and use Fess Site Search as a free replacement for Google Site Search, generating the embed through the FSS JS Generator.
  • Teams that must gate results by identity, using role- and permission-based filtering together with SAML, OpenID Connect, SPNEGO, or Microsoft Entra ID single sign-on.
  • Multilingual organizations whose content spans more than one language, relying on the 20+ language text analysis support.
  • Administrators who prefer browser configuration to index-level work: crawling targets are registered on the Web, File, or Data Store configuration pages, then the crawler is started from the Scheduler page.

Getting started

Download a DEB, RPM, or ZIP package from the Releases page and run it, for example by unzipping the ZIP and executing ./bin/fess, or run the published Docker images from ghcr.io with the Docker Compose file in the docker-fess repository. Java 21 or later is required for the ZIP, RPM, and DEB packages, and the search UI is then served at http://localhost:8080/ with the admin UI at http://localhost:8080/admin/.

How it compares

Fess does not compete with OpenSearch or Elasticsearch so much as sit on top of them: it supplies crawling, administration, permissions, and a search front end that a raw Lucene-based engine does not. Its closest named comparison point is Google Site Search, which the companion Fess Site Search project replaces as a free, self-hosted option. Where Elasticsearch and OpenSearch are general-purpose engines, Fess is the assembled enterprise search server built around one of them.

When to use it — and when not to

A self-hoster must run a Java 21 runtime and, for non-Docker installations, a separate OpenSearch backend, since only the Docker images bundle it. Anyone who wants a fully managed, vendor-hosted search service should look elsewhere, because Fess is designed for self-hosting and browser-based administration. The README emphasizes crawling, formats, and authentication far more than the AI, RAG, and semantic search capabilities that appear in the project tagline and topics, so buyers should verify that side of the feature set before committing.

project readme (upstream, from github) — read inline

Fess: Enterprise Search Server

Java CI with Maven License GitHub Release

Overview

Fess is an enterprise search server that you can install and run on any platform with a Java runtime. It is built on OpenSearch, but prior OpenSearch knowledge is not required: Fess is configured through a browser-based administration UI.

A built-in crawler collects documents from web sites, file systems, and data stores such as databases and CSV files. Many file formats are supported, including Microsoft Office, PDF, and ZIP archives.

Fess Site Search is a free alternative to Google Site Search that you can embed in your own website. For details, see the FSS JS Generator documentation.

Features

  • Full-text search with faceting, sorting, and search suggestions
  • Crawlers for web sites, file systems, and data stores (databases, cloud storage, and SaaS)
  • Support for many document formats, including Microsoft Office, PDF, and archives
  • Browser-based administration UI and a REST API
  • Role- and permission-based filtering of search results
  • Single sign-on with LDAP, OpenID Connect, SAML, SPNEGO, and Microsoft Entra ID
  • Multilingual user interface and text analysis (20+ languages)
  • Extensible through data store, ingest, script, and theme plugins

Requirements

  • Java 21 or later, for the ZIP, RPM, and DEB packages
  • OpenSearch as the search engine backend. The Docker images bundle it; for other installations you set it up separately.

See the Installation Guide for supported versions and setup details.

Getting Started

There are two ways to try Fess: download and install it yourself, or run it with Docker.

Download and Install/Run

Fess can be downloaded from the Releases page. Downloads are available in three formats: DEB, RPM, and ZIP.

The following commands show how to use the ZIP download:

$ unzip fess-.zip
$ cd fess-
$ ./bin/fess

For more details, see the Installation Guide.

Docker

Docker images are published on ghcr.io. A Docker Compose file is available in the docker-fess repository.

Browser UI

Search UI

Admin UI

Register crawling targets on the Web, File, or Data Store configuration pages in the Admin UI, then start the crawler from the Scheduler page.

Migration from Another Search Provider

See MIGRATION.md.

Data Store

Fess can crawl the following storage locations and APIs:

Plugins

Fess can be extended with plugins. In addition to the data store connectors above, the following are available.

Theme

  • fess-themes: a collection of static themes, each a self-contained single-page app installed by uploading a ZIP in the admin UI

Ingest

Script

Development

Get the Source Code

  1. Clone the repository:

    $ git clone https://github.com/codelibs/fess.git
    
  2. Import it as a Maven project in your IDE.

Building Fess requires Java 21 or later and Maven.

Set Up OpenSearch Plugins

Run antrun:run to download plugins into the plugins directory:

$ mvn antrun:run

Run Fess

Run or debug org.codelibs.fess.FessBoot in your IDE, then open http://localhost:8080/.

Build a Package

Run the package goal to create the release file in target/releases:

$ mvn package
$ mvn rpm:rpm   # .rpm package
$ mvn jdeb:jdeb # .deb package

Generate Source Code

$ mvn dbflute:download # one-time setup
$ mvn dbflute:freegen
$ mvn license:format

Integration Tests

Integration tests require a running Fess server with OpenSearch.

  1. Build Fess:

     $ mvn antrun:run  # download OpenSearch plugins, if not done already
     $ mvn package
    
  2. Start the Fess server:

     $ unzip target/releases/fess-*.zip
     $ ./fess-*/bin/fess &
    

    Wait for Fess to become ready (this may take up to 60 seconds). It returns a JSON response when ready:

     $ curl -s "http://localhost:8080/api/v1/health"
    
  3. Clone the test data (required for SearchApiTests):

     $ git clone https://github.com/codelibs/fess-testdata.git /tmp/fess-testdata
    
  4. Run the tests:

     $ mvn test -P integrationTests -Dtest.fess.url="http://localhost:8080" -Dtest.search_engine.url="http://localhost:9201"
    

    To run a single test case:

     $ mvn test -P integrationTests -Dtest.fess.url="http://localhost:8080" -Dtest.search_engine.url="http://localhost:9201" -Dtest=SearchApiTests
    

Contributing

Contributions are welcome. If you find a bug or have a feature request, please open an issue. For questions and discussion, use the forum. Pull requests for fixes and improvements are appreciated; for larger changes, discussing the approach first is recommended.

Before committing, format the code and license headers:

$ mvn formatter:format
$ mvn license:format

Adding a Language

Fess is internationalized. To add labels and messages for a language, translate the property files below and rename them with your language code (fess_*_[lang].properties):

For search and index analysis, if doc.json contains lang_[lang] for your language, adjust the analyzer accordingly. See the OpenSearch documentation on analyzers for details. Pull requests for new languages are welcome.

Community and Support

Translations

This README is available in other languages:

Powered By

License

Fess is distributed under the Apache License 2.0.

Frequently asked questions

Is fess free to use?

fess is open source under the Apache-2.0 licence. There is no licence fee and no seat count — you can self-host it or, where the project offers one, pay a vendor for a managed version instead.

What does fess do?

Open-source, self-hosted enterprise & site search server built on OpenSearch. Crawls web / file / DB / cloud sources, 20+ languages, REST API, and AI/RAG & sema

What is fess written in?

fess is primarily written in Java. Its source is publicly available at https://github.com/codelibs/fess, and it has 1,142 GitHub stars.