The two components have one prerequisite in common: they both require Java to run. While this is the only requirement for the REST server, the Joex components requires some more external programs.
The rest server and joex components are not required to "see" each other, though it is recommended.
Very often, Java is already installed. You can check this by opening a
terminal and typing
java -version. Otherwise install Java using your
package manager or see this site for
It is enough to install the JRE. The JDK is required, if you want to build docspell from source. For newer versions, the JRE is not shipped anymore, simply use JDK then.
Docspell has been tested with Java 11 (or sometimes referred to as JDK 11). The provided packages are build using JDK 11. However, it also works on newer java versions. The provided docker images use JDK11.
The next tools are only required on machines running the Joex component.
gscommand) is used to extract/convert PDF files into images that are then fed to ocr. It is available on most GNU/Linux distributions.
The performance of
unoconv can be improved by starting
in a separate process. This runs a libreoffice/openoffice listener and
therefore avoids starting one each time
unoconv is called.
On Debian this should install all joex requirements:
sudo apt-get install ghostscript tesseract-ocr tesseract-ocr-deu tesseract-ocr-eng unpaper unoconv wkhtmltopdf ocrmypdf
SOLR is used to provide the fulltext search feature. This feature can be disabled, so installing SOLR is optional. But without it, there is no fulltext search.
When installing manually (i.e. not via docker), just install solr and create a core as described in the solr documentation. That will provide you with the connection url (the last part is the core name).
When using the provided
docker-compose.yml setup, SOLR is already setup.
SOLR must be reachable from all joex and all rest server components.
Both components must have access to a SQL database. The SQL database contains all data (including binary files) and is the central component of docspell. Docspell has support these databases:
The H2 database is an interesting option for personal and mid-size setups, as it requires no additional work. It is integrated into docspell and works really well out of the box. It is also configured as the default database.
When using H2, make sure that all components access the same database
– the jdbc url must point to the same file. Then, it is important to
add the options
;MODE=PostgreSQL;DATABASE_TO_LOWER=TRUE;AUTO_SERVER=TRUE at the end
of the url. See the config page for
For large installations, PostgreSQL or MariaDB is recommended. Create a database and a user with enough privileges (read, write, create table) to that database.