HumanSign is an open-source project that explores keystroke dynamics as a method for human verification and digital notarization.
The system analyzes how a user types (timing, rhythm, patterns) rather than what they type, to help verify identity in a non-intrusive way.
Keystroke dynamics is a behavioral biometric technique that identifies users based on typing behavior such as:
- Key press and release timing
- Typing rhythm and patterns
- Consistency across sessions
HumanSign uses these characteristics to help distinguish genuine human users from impersonation or automated behavior.
The goal of HumanSign is to:
- Build a transparent and auditable keystroke-based verification system
- Explore its use as a notary-style trust mechanism
- Promote open research and ethical use of behavioral biometrics
| The Problem | The Solution |
|---|---|
| Current AI-detection systems rely on content analysis, which can be easily bypassed using "AI humanizer" tools. | HumanSign ignores content and focuses on the writing process (keystroke dynamics) for verification. |
| Honest work is falsely flagged as AI-generated, especially from non-native speakers, while sophisticated AI content passes undetected. | We provide cryptographic proof of authorship and authenticity through typing behavior and patterns. |
| Adversarial AI now reliably beats all traditional statistical content detectors. | Our solution uses behavioral biometrics (typing rhythm and timing), which are harder to fake or manipulate than content alone. |
- Identity verification
- Continuous authentication
- Anti-bot or anti-impersonation systems
- Research and experimentation with behavioral biometrics
This project is under active development and primarily intended for:
- Academic exploration
- Prototyping
- Open-source collaboration
It is not production-ready and should not be used as a sole security mechanism.
This project is licensed under the GNU Affero General Public License v3.0 (AGPL-3.0). Any modifications or network-based use must make the source code available.
Contributions, ideas, and discussions are welcome.
Please open an issue or pull request to get involved.
HumanSign is a research-oriented project.
Behavioral biometric systems may have ethical, privacy, and accuracy limitations. Use responsibly.