Skip to content
Switch branches/tags

Latest commit


Git stats


Failed to load latest commit information.
Latest commit message
Commit time

Nanopub Server

This is an implementation of a nanopublication server. Such servers form a server network, which can be used to publish nanopublications that have trusty URIs. (Such a server only returns entire nanopubs. No queries supported; no triple store involved.)

Nanopublications are tiny snippets of data with provenance information attached. They can be organized in a very flexible manner into large datasets (which can also be described as nanopublications). More details can be found in the publications listed below.

To access the server network, the nanopub-java library can be used.


The following publications explain the approach:

  • Kuhn et al. Publishing without Publishers: a Decentralized Approach to Dissemination, Retrieval, and Archiving of Data. In Proceedings of the 14th International Semantic Web Conference (ISWC) 2015.
  • Kuhn et al. Decentralized provenance-aware publishing with nanopublications. PeerJ Preprints.
  • Kuhn. nanopub-java: A Java Library for Nanopublications. In Proceedings of the 5th Workshop on Linked Science (LISC 2015).

Current Network

Have a look at this nanopub monitor (source code) to see the current nanopub server network:

From that monitor page, you can find the URLs of the current servers in the network, for example this one:


Easy Deployment with Docker


  • Docker 1.10 or higher
  • Docker-compose 1.6 or higher

After cloning the repository, execute:

$ docker-compose up

Config can be changed by passing environment variables prefixed with NPS_. E.g. setting is done via variable NPS_MONGODB_HOST.

Manual Compilation and Deployment


  • Java 1.7 or higher
  • Access to a MongoDB instance
  • Disk space of up to around 15kB per nanopublication (10-11kB is average so far, but this value might change; the current size of the network can be seen here:

In addition, we assume here that Git and Maven are installed (tested with Maven version 3; version 2.0.9 might work too).

Now you can fetch the code for the nanopub server:

$ git clone

To configure the server, make a copy of the configuration file with the prefix local., which overrides the main configuration file:

$ cd nanopub-server/src/main/resources/ch/tkuhn/nanopub/server/
$ cp

Edit the file to configure your instance of the server, and then compile it (run in the top nanopub-server directory):

$ mvn package

(Depending on the state of this repository, you might also have to install the latest snapshot version of with git clone ... and mvn install.)

Then you can run the nanopub server using Maven's Jetty plugin:

$ mvn jetty:run

Now you should be able to locally access the server from your browser:

Alternatively, you can give the file target/nanopub-server.war to a web application server such as Apache Tomcat. Then you can setup your web server to map a public URL to the nanopub server, for example: >

Add the public URL to the line public-url= of the configuration file and recompile and restart the server. Then, it will connect to the server network.


Usage is pretty straightforward when you look at the web interface that a nanopub server provides. To retrieve a nanopub like

the artifact code (last 45 characters) has to be given to the nanopub server like this:

File extensions can be used to retrieve the nanopub in different formats:

Depending on your browser settings, these URLs might open in an external editor or you might have to download them. In order to view them directly in the browser, .txt can be appended:


This nanopub server is free software under the MIT License. See LICENSE.txt.


A simple server to publish nanopublications




No packages published