Cerno is an open-source TypeScript SDK designed for human verification without relying on hardware. It addresses the challenge of distinguishing human users from automated agents by analyzing motor-control behavior during maze interactions. The tool operates through a sequence of steps, including proof-of-work, maze generation, behavioral analysis, Stroop probes, signature binding, and reputation tracking, all aimed at robustly verifying human presence in a privacy-conscious manner.
The verification process begins with a proof-of-work step, where clients compute a SHA-256 hash with adaptive difficulty based on client signals. Mazes are generated using the Growing Tree algorithm with a seeded pseudorandom number generator, enabling the server to regenerate mazes from the same seed for trustless validation. During interaction with the maze, Cerno extracts twelve behavioral features from raw pointer events, seven of which are public and five server-only. These features include measures such as standard deviation of tangential velocity, path efficiency, pause count, movement onset time, jerk standard deviation, angular velocity entropy, and coefficient of variation of inter-event timing. The server scores these features against per-maze baselines to assess authenticity.
Additional verification layers include Stroop probes, where color-word interference tasks are presented at maze decision points, with the server deriving response timings from the event stream. Signature binding is implemented using an ephemeral ECDSA P-256 keypair, binding the public key at challenge issuance and verifying it upon submission. Reputation tracking is handled through behavioral consistency analysis across sessions, using exponentially weighted moving average trust scores keyed by a stable device identifier.
Cerno is delivered as a TypeScript SDK, with integration examples provided for both client and server environments via npm packages (@cernosh/react and @cernosh/server). 0 license, and its open-source codebase is available on GitHub. The tool is suitable for developers seeking a privacy-friendly, hardware-free solution to human verification challenges.
Cerno sits in PulseGate's Frameworks & SDKs category. It focuses on verifying human users online without hardware or third-party ML, using behavioral analysis of maze interactions. It is built as an open-source project for web developers. Cerno is open source under the Open Source license. Cerno is available on the web and the command line, and it can be self-hosted.
Cerno Maintainers builds and maintains Cerno, and the product first shipped in 2026. Development happens publicly on GitHub with 62 commits in the last 90 days. PulseGate's similarity index finds few close equivalents — Cerno occupies a relatively distinct niche. Key capabilities include maze interaction, motor-control analysis, and behavioral scoring.
Latest indexed changes and source events
Other apps tracked under the same category.
