A nanopublication is a small knowledge graph snippet with metadata that is treated as an independent (scientific) publication.
In its own words: https://nanopub.net/, read 2026-10-29
Its own pages call nanopublications "FAIR data containers for scientific results, and more," whose content may be "about anything," a relation between a gene and a disease or an opinion, standing as independent publications "with or without an accompanying research article." Other formulations call one "the smallest unit of publishable information," an assertion "uniquely identified and attributed to its author."
From its /non article, sentences 2, 3
The model was "independently conceived" from 2009 as "a minimal, uniform container making individual semantic assertions and their associated provenance metadata, machine-actionable." Barend Mons and Jan Velterop are credited with proposing it for single, attributable and machine-readable assertions in the scientific literature.
From its /non article, sentence 11
Each nanopublication has three basic elements: the assertion, its main content as a small atomic unit of information; the provenance, which "describes how the assertion above came to be," its methods and the kind of study performed; and the publication info, "when and by whom it was created and the license terms for its reuse." One account gives four parts, a head joining the three graphs.
From its /non article, sentences 6, 8
The carrier is RDF. Identity rests on resolvable Trusty URIs, and signing and publishing run over HTTP. The core services are second-generation: the Nanopub Registry publishes and retrieves nanopublications in a decentralized fashion, Nanopub Query queries them through SPARQL endpoints, and the legacy nanopub-server keeps the old unsigned ones. Its tools include the Nanodash client, the nanopub-py Python client and a nanopub Agent Skill for AI coding agents.
From its /non article, sentences 9, 13, 14, 15
The first practical application, in 2014, published millions of nanopublications for the FANTOM5 Project, and more than ten million were counted in 2018. Attribution is the stated incentive: "because nanopublications can be attributed and cited, they provide incentives for researchers to make their data available in standard formats." A paper of 2021 found them "still not widely used by scientists outside specific circles; they are hard to find and rarely cited."
From its /non article, sentences 17, 19, 20
The scope is the assertional data type, each nanopublication limited to a single assertion or a small RDF graph. The model expresses and records provenance and does not measure its survival, and publication and signing do not settle an assertion's truth.
From its /non article, sentences 21, 22
The archive reads the paradigm as treating "machine-readability primarily as data architecture," while machine reception "remains sensitive to compositional properties that no schema captures," among them the "proximity of provenance to claims." Its Claim Status Packet follows nanopublication practice, keeping assertion, provenance and publication information as separate graphs, and adds the question of what standing a claim is permitted to carry.
From its /non article, sentences 24, 27
At the bare name, Google's AI Overview composed the model from nanopub.net's own definition and three parts, named no person, institution or identifier, and left out signing, the second-generation Registry and Query, and the legacy server.
From its /non article, sentences 7, 29
This article is derived from the entity's /non article (EA-NEGONT-02, the knowledge object composed under v0.9): every fact in it is stated there, and each paragraph names the sentences it rests on.
| What receives provenance | an assertion: a small knowledge-graph snippet published on its own |
| Relations carried | assertion provenance (how the assertion came to be); publication info (who created the nanopublication, when, licence); signature (signing at publication) op |
| Carrier | RDF graphs: assertion, provenance and publication-info parts of one nanopublication op |
| Expresses · records · tests | expressesrecordstests |
| Operations tested | none |
| What survived | not tested: the model carries provenance; it does not measure its survival |
op the operator's own statement · 3p a third-party account · obs our observation · rd our reading. Hover a tag for its quote; each opens below with its source.
| Canonical name | Nanopublications |
| Also known as | nanopub, Nanopub Registry, Nanopub Query, Nanodash |
| Address | https://nanopub.net/ |
| Operator | the nanopublication community (nanopub.net) |
| Kind | publication model and service ecosystem |
| Unit | a nanopublication |
| Lifecycle | active — second-generation Registry and Query services; the legacy nanopub-server keeps old unsigned nanopublications op |
| Data access | decentralized publishing and SPARQL query services op |
| Versions and dates | architecture page updated: 2026-10-28; read: 2026-10-29 |
| Observation | 2026-10-29: read |
b2d9af40755a66ed…b2d9af40755a66ed…b2d9af40755a66ed…b2d9af40755a66ed…0e21d7ae559d8301…0e21d7ae559d8301…0e21d7ae559d8301…b2d9af40755a66ed…0e21d7ae559d8301…Lee Sharks, ed. "Nanopublications." Provenance meta-aggregator, Provenance Erasure, compiled 2026-10-29. https://www.provenanceerasure.org/aggregators/nanopublications
Record https://www.provenanceerasure.org/aggregators/nanopublications · entity https://www.provenanceerasure.org/aggregators/nanopublications#entity · the census as data
@misc{pe-nanopublications2026,
editor = {Sharks, Lee},
title = {Nanopublications},
howpublished = {Provenance meta-aggregator, Provenance Erasure},
year = {2026},
note = {Compiled 2026-10-29; every field sourced and dated},
url = {https://www.provenanceerasure.org/aggregators/nanopublications}
}