A behavioural authentication and network traffic logger for game environments.
BEACON-Logger is the data acquisition tool behind the BEACON dataset. It runs in the background during a gameplay session and records five synchronised streams: keystroke dynamics, mouse kinematics, network traffic, screen video and hardware metadata. It also captures the game's local configuration, so that sensitivity and keybind differences can be accounted for in behavioural models.
It was built for research on continuous authentication and behavioural biometrics, where the question is whether the way a person uses their computer can identify them reliably for as long as they are at the machine.
| Date | Milestone |
|---|---|
| Jan 2025 | Logger development begins |
| Apr 2025 | Development of the backend that receives participant uploads begins |
| Jun 2025 | Participant upload portal development begins |
| 2025 | Six-month data collection campaign in Lab L301, Thapar Institute of Engineering & Technology |
| Aug 2025 | Final changes to the logger before release |
| 29 Apr 2026 | BEACON dataset published on Hugging Face |
| 7 May 2026 | v1.0.0 released on Zenodo (10.5281/zenodo.20062628) |
| 11 May 2026 | Paper published on arXiv (2605.10867) |
| 13 May 2026 | v2.0.0 released on Zenodo (10.5281/zenodo.20144661) |
The Git history of this repository mirrors the Zenodo releases, with one commit and tag per release, dated to its publication. Zenodo remains the archival record.
| Stream | Output | Details |
|---|---|---|
| Keyboard | keyboard_<user>_<time>.csv |
Press and release times, hold duration, inter-key latency, concurrent keys |
| Mouse | mouse_<user>_<time>.csv |
X/Y position, speed, acceleration, clicks, double clicks, scrolls |
| Network | captured_packets_<user>_<time>.pcap |
Raw packets captured with Scapy and Npcap |
| Screen | screen_record_<user>_<time>.mp4 |
Session video through FFmpeg and a DirectShow capture filter |
| Hardware | hardware_info_<user>_<time>.json |
OS, CPU, RAM, monitors and connected HID devices through WMI |
| Game config | valorant_config_<user>_<time>/ |
Copy of the local VALORANT configuration folder |
Every session is written to its own timestamped folder, data_YYYY-MM-DD_HH-MM-SS/.
beacon-logger/
├── logger.py the logger, exactly as released on Zenodo
├── requirements.txt Python dependencies
├── docs/
│ └── README_logger.md the original release documentation, unmodified
├── assets/
│ └── logger.png screenshot of the logger running
├── CHANGELOG.md
├── CITATION.cff
└── LICENSE CC BY-NC-ND 4.0
The logger targets Windows.
A packaged logger.exe is published with
v2.0.0 on Zenodo. It is not stored in this
repository because it exceeds GitHub's file size limit.
Prerequisites:
- Python 3
- Npcap for packet capture
- FFmpeg, placed so that
ffmpeg/bin/ffmpeg.exesits next tologger.py - Screen Capturer Recorder for video capture
pip install -r requirements.txt
python logger.pyThe logger asks for the participant's name and explicit consent before it records anything. Press Left Ctrl + Esc together to stop a session.
This tool records keystrokes, screen video and network traffic. Run it only on machines you are authorised to monitor, and only with the informed consent of the person using them. It was developed strictly for research, and the authors accept no responsibility for misuse. The public BEACON dataset was anonymised before release, as described on its dataset card.
- Paper: BEACON: A Multimodal Dataset for Learning Behavioral Fingerprints from Gameplay Data
- Dataset: beacon-gui/BEACON-Dataset on Hugging Face, and 10.5281/zenodo.20034625 on Zenodo
- Software: 10.5281/zenodo.20144661 on Zenodo
If you use this tool, please cite the paper and the software.
@misc{singh2026beaconmultimodaldatasetlearning,
title = {BEACON: A Multimodal Dataset for Learning Behavioral Fingerprints from Gameplay Data},
author = {Ishpuneet Singh and Gursmeep Kaur and Uday Pratap Singh Atwal and Guramrit Singh and Gurjot Singh and Maninder Singh},
year = {2026},
eprint = {2605.10867},
archivePrefix = {arXiv},
primaryClass = {cs.CR},
url = {https://arxiv.org/abs/2605.10867}
}
@software{singh2026beaconlogger,
title = {BEACON-Logger: A Behavioral Authentication and Network Traffic Logger for Game Environments},
author = {Ishpuneet Singh and Guramrit Singh and Gursmeep Kaur and Uday Pratap Singh Atwal and Gurjot Singh and Maninder Singh},
year = {2026},
version = {v2},
publisher = {Zenodo},
doi = {10.5281/zenodo.20144661},
url = {https://doi.org/10.5281/zenodo.20144661}
}GitHub also reads CITATION.cff and offers these citations from the
repository sidebar.
Released under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International licence, matching the v2.0.0 Zenodo record. v1.0.0 was released under CC BY 4.0.
Ishpuneet Singh, Guramrit Singh, Gursmeep Kaur, Uday Pratap Singh Atwal, Gurjot Singh and Maninder Singh, at the Thapar Institute of Engineering & Technology.
We thank Dr. Maninder Singh (CSED, TIET) and Gurjot Singh for their mentorship, and Akshat Kaushik, Ishika Walia and Rudraksh Chugh for monitoring the collection lab for most of the six-month campaign. We also thank the Thapar Institute of Engineering & Technology for the lab space and infrastructure, and every participant who contributed their gameplay.