Repolode
Submit
fabio-rovai

fabio-rovai/open-ontologies

A Rust engine for planning, applying and certifying changes to OWL/RDF ontologies, with blast-radius reports and checkable Lean 4 proofs of derived conclusions.

Open-source alternative to:Protégé

View on GitHubWebsite
Stars
1,159
Forks
161
Open issues
17
Last commit
3 days ago

What is it?

Open Ontologies is a Rust command-line tool and MCP server for working on OWL and RDFS ontologies. Before a change reaches production, it computes a semantic plan that reports added and removed classes, the triples affected, a risk score, and the new consequences the change would entail. A text diff cannot show those consequences, so the plan is aimed at changes that look correct line by line but infer extra facts.

Reasoning runs can write a derivation certificate: two plain TSV files listing asserted triples and each inferred step with its rule and premises. A separate Lean 4 checker, built without Mathlib, re-verifies that certificate without needing this engine or a network connection. The checker also rejects forged conclusions and reports the rule that fails. Other features include SHACL validation, SPARQL 1.1 through Oxigraph, loading CSV, JSON, XML, YAML, XLSX, Parquet, PostgreSQL and DuckDB data as RDF, and export to TPTP, CLIF, SMT-LIB and LADR.

The project ships as one binary with prebuilt releases, a Docker image, and an MCP server for clients such as Claude and Cursor.

Key features

  • 01Plans ontology changes with added and removed classes, blast radius and a risk score before applying them.
  • 02Reports new consequences that a change would entail, beyond what a text diff shows.
  • 03Writes derivation certificates that a Lean 4 checker re-verifies without the engine or network access.
  • 04Rejects forged conclusions and names the rule that fails.
  • 05Validates data against SHACL and queries RDF with SPARQL 1.1.
  • 06Loads CSV, JSON, XML, YAML, XLSX, Parquet, PostgreSQL and DuckDB data into RDF.

Use cases

  • Reviewing a change to a production ontology before it ships.
  • Giving an auditor a certificate they can re-check months later without installing the software.
  • Connecting an AI assistant such as Claude or Cursor to ontology tooling through MCP.
  • Converting tabular or relational data into RDF for semantic integration.

Quick start

  1. 1

    Build the Rust binary from the repository

  2. 2

    Build the Lean 4 checker

    cd lean && lake build
  3. 3

    Load a small ontology and run RDFS reasoning with a certificate

    open-ontologies --data-dir /tmp/oo-demo/store reason --profile rdfs --certificate ./cert

Things to know

  • The README does not give a package-manager install command, so setup needs a manual build.
  • Unsatisfiability answers carry no guarantee, as the README itself states.
  • Full value depends on running the separate Lean 4 checker, which is an extra setup step.

Topics

#ontology#owl#rdf#reasoning#sparql#shacl#rust#formal-verification#knowledge-graph#ai-native

FAQ

What is open-ontologies?
A Rust engine for planning, applying and certifying changes to OWL/RDF ontologies, with blast-radius reports and checkable Lean 4 proofs of derived conclusions. Open Ontologies is a Rust command-line tool and MCP server for working on OWL and RDFS ontologies. Before a change reaches production, it computes a semantic plan that reports added and removed classes, the triples affected, a risk score, and the new consequences the change would entail. A text diff cannot show those consequences, so the plan is aimed at changes that look correct line by line but infer extra facts.
Is open-ontologies free to use? What license does it have?
open-ontologies is open source under the MIT license. Check the license terms before using it in commercial products.
Is open-ontologies still maintained?
Actively maintained. The last commit to open-ontologies was on Oct 17, 2026.
What can I use open-ontologies for?
Reviewing a change to a production ontology before it ships. Giving an auditor a certificate they can re-check months later without installing the software. Connecting an AI assistant such as Claude or Cursor to ontology tooling through MCP. Converting tabular or relational data into RDF for semantic integration.
How do I install open-ontologies?
Build the Rust binary from the repository Build the Lean 4 checker: cd lean && lake build Load a small ontology and run RDFS reasoning with a certificate: open-ontologies --data-dir /tmp/oo-demo/store reason --profile rdfs --certificate ./cert
Where can I get open-ontologies?
The source code and installation instructions are on GitHub: https://github.com/fabio-rovai/open-ontologies

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects