AGNA

Who is AGNA for?

AGNA is particularly suited to small and medium research networks: communication, collaboration, affiliation, interaction, behavioral, institutional, historical, and other relational data. It can be used in sociology, anthropology, communication research, psychology, education, organizational research, animal-behavior studies, and the humanities.

At a glance

Source, licensing, and reproducible workflows

AGNA Desktop, AGNA Core, and AGNA CLI are free and open-source software released under the Apache License, Version 2.0. The licence permits use, modification, distribution, and commercial use, subject to preservation of applicable copyright, licence, and notice material. It also includes an explicit patent licence. Read the full Apache License 2.0.

The projects share a common architecture: AGNA Desktop provides the graphical interface; AGNA Core provides the reusable engine; and AGNA CLI makes core functions available for command-line, scripted, and reproducible work.

AGNA Desktop  ──────►  AGNA Core  ◄──────  AGNA CLI
(graphical UI)          (engine)           (command line)

Browse source code · Read the contribution guide · Report a bug or propose an improvement · Report security vulnerabilities privately

Build from source

git clone https://github.com/baeckett/AGNA.git
cd AGNA
mvn verify   # requires JDK 17 and Maven

The repository builds with Maven against JDK 17; the full test suite (230 tests) runs with mvn verify.

Use AGNA Core

AGNA Core is available as a release JAR. Browse the AGNA Core module for source and build information. Generate API documentation locally with mvn javadoc:javadoc.

Installation

Download AGNA

The desktop installers bundle the Java runtime, so Java is not necessary to be installed separately. The standalone native CLI binaries do not require Java at all (sub-10ms startup, single executable). The plain cross-platform desktop JAR requires Java 17 or newer.

Install AGNA Desktop

System requirements

• Supported systems: AGNA runs on Windows, macOS, and Linux. The native CLI binaries run directly with zero dependencies. The plain desktop JAR requires Java 17 or newer; the GUI installers bundle a compatible Java runtime.
• Memory: 512 MB minimum (native CLI operates under 15 MB RSS); 2 GB recommended for large desktop graphs.
• Disk space: about 150–250 MB installed for Desktop; ~95 MB for standalone native CLI.
• Internet connection: not required for normal analysis after installation.

Advanced & developer downloads

Read the full AGNA 2.2.1 release notes and verify downloads with SHA-256 checksums.

First steps

Quick Start

This short exercise introduces a small directed communication network. It is intended to help a new user open data, inspect a graph, calculate basic measures, and save results. See the full Quick Start guide.

1. Download the example data

Download the six-person communication example. The file should contain either an adjacency matrix or an edge list, depending on the version of AGNA you release.

2. Open the network

1. Start AGNA Desktop.
2. Select File → Open.
3. Choose the example file.
4. Confirm whether ties are directed, and whether values are binary or weighted.

3. Inspect the network

Open the graph view. Each node represents a person, and each arrow represents a directed communication tie. Drag a node to improve readability if necessary. Notice which people receive or send several ties and whether any are isolated.

4. Calculate basic measures

1. Run the network description command to obtain the number of nodes, ties, density, and components.
2. Run degree centrality to inspect who has the most incoming or outgoing relations.
3. Run shortest-path or distance analysis to examine how easily actors can reach each other through the network.

5. Save and export

1. Save the AGNA project under a new name.
2. Export any table needed for later work.
3. Export the graph image or vector graphic for a presentation or paper.
4. Record the AGNA version and the analytical choices used.

Continue with the documentation for importing data, working with different tie types, interpreting measures, and exporting results.

Screenshots

Screenshot of the AGNA Desktop application on MacOS Tahoma:

Manual and learning resources

Documentation

AGNA documentation is organized around research tasks and reproducible workflows.

Examples and teaching materials

Browse example networks and teaching exercises. Each example should include a small data file, a research question, expected output, and interpretive notes.

Need help?

See the support page, search the issue tracker, or report a reproducible problem using the information requested there. Please remove confidential network data before attaching example files to a public issue.

Research use

Cite AGNA

If you use AGNA in research, teaching, a thesis, or a publication, please cite the software version used.

Preferred software citation

Bența, M. I. (2026). AGNA: Applied Graph and Network Analysis Open Source
(Version 2.2.1) [Computer software]. AGNA Open Source.
https://doi.org/10.5281/zenodo.24226451

BibTeX

@software{benta_2026_agna,
  author    = {Bența, Marius I.},
  title     = {{AGNA}: Applied Graph and Network Analysis Open Source},
  version   = {2.2.1},
  year      = {2026},
  publisher = {AGNA Open Source},
  doi       = {10.5281/zenodo.24226451},
  url       = {https://github.com/baeckett/AGNA}
}

View the BibTeX citation file · View the Citation File Format record · Zenodo deposit (Concept DOI: 10.5281/zenodo.22708198)

Historical paper

For historical background on AGNA 2.1, see Studying Communication Networks with AGNA 2.1.

Version history

Releases

Every release includes a version number, release notes, source tag, downloadable artifacts, and checksums. Record the version you use in research outputs.

Known issues

No critical known issues are currently listed. See the release notes and issue tracker for current limitations.

Verify downloads

Advanced users can compare the SHA-256 checksum of a downloaded installer, archive, or native binary with the values published for the release. This helps confirm that the file matches the released artifact.

# Example only — adapt to the user’s operating system
sha256sum agna-linux-amd64.tar.gz

Project history

About AGNA

AGNA—Applied Graph and Network Analysis Open Source—was developed as a free application for social network analysis, sociometry, and sequential analysis by Marius Ion Bența, currently a Sociology Researcher with George Barițiu History Institute and Guest Lecturer with the Babeș-Bolyai University in Cluj.

From its beginning, AGNA was intended for researchers, teachers, and students who needed accessible network editing and vizualization tools rather than a programming environment. Ease of use, direct visual interaction, and clarity of output were central priorities.

AGNA today

AGNA 2.2.1 is the result of an independent project and is a stable open-source release. It preserves the Desktop application while separating the reusable computational engine as AGNA Core and providing AGNA CLI with standalone zero-runtime native binaries and MCP support for reproducible command-line and AI agent workflows.

The project’s long-term aim is reasonable and practical: to keep accessible network analysis available to researchers who work with relational data and to provide an open foundation for teaching, scholarship, and future development.

With limited resources, development and support are offered on a best-effort basis, and constructive feedback, bug reports, and contributions are warmly welcome.

Acknowledgements

The author gratefully acknowledges the researchers, teachers, students, and early users whose feedback helped shape AGNA’s design and its emphasis on accessibility for non-programming social scientists and humanists.

The author also extends many thanks to Prof. Alina S. Rusu for her research and educational collaboration in the AGNA project.

Third-party software, libraries, icons, fonts, and other materials are listed separately in THIRD_PARTY_NOTICES.md.

The story of AGNA

Marius, the author of AGNA, started working on this software in the 90s as a student in sociology in Cluj city, as he was learning about social network analysis and sociometry from Professor Achim Mihu and from the works of Claude Flament and Alex Bavelas. Initially, there were just code snippets written in Turbo Pascal that would perform computations related to social network analysis and sequential analysis. Later, they turned into a desktop application developed in Delphi that could run on Windows. Around the year 2000, the software was rewritten in Java, given the write-once-run-anywhere promise of this programming language, and was distributed as a free desktop application. Over the years, AGNA was used in hundreds of research and education projects worldwide. However, given that this was an independent project, around the year 2005 maintenance had to be halted as the author had to be involved in other activities and jobs.

In 2026, the code was revised, which led to significantly better peformance and stability. On September 14^th 2026, AGNA was published as an open source project on GitHub under the Apache 2.0 Licence. AGNA was released in the form of three separate components—a desktop application that can be installed on any operating system, a core library that can be integrated in other software, and a command-line interface for those who prefer performing automated computations without a graphical user interface—with the hope that this new configuration and this new version would be more useful and would better suit the needs of the scientific community in research and teaching.

The author is grateful to everyone from the academic world, from business, and from various institutions who have used AGNA in their projects, to all the people who took their time to provide precious comments and feedback, and to God for making this possible.

Contact information

• General contact: contact@netanalysis.co.uk
• Source code and issue tracker: https://github.com/baeckett/AGNA
• Private security reports: contact@netanalysis.co.uk

Privacy and security

AGNA Desktop and AGNA CLI process network data locally on the user’s computer. They do not upload project files, node labels, edge information, or analysis results to AGNA servers.

For ordinary questions, documentation corrections, feature suggestions, and reproducible bug reports, please use the AGNA issue tracker. Please do not include confidential research data in public issue reports. To report a potential security vulnerability privately, please contact contact@netanalysis.co.uk before disclosing it publicly.