No AI can replicate how you type. Authored certifies genuine human authorship through keystroke biometrics — without ever reading a single word you write. A verifiable proof of humanity built for institutions, employers and publishers.
No spam. Unsubscribe anytime. We'll email you when Authored launches.
Keystroke dynamics
124ms
Avg Dwell
8.3%
Backspace
62
WPM
The scale of the problem Authored solves
+0%
increase in academic cheating in 12 months
ICAI 2024
<0%
accuracy of AI detectors against humanizers
Stanford, 2024
>0%
target accuracy in human authorship verification
Authored Research
0
false positives — by design, not by tuning
Authored Architecture
The problem
Content-based detectors analyze what was written — making them inherently vulnerable to humanizers and paraphrasers. They generate false positives that wrongfully flag genuine human work.
Keystroke rhythm — human vs AI/pasted content. The difference is unmistakable.
How it works
From first keystroke to verifiable certificate — the entire process is invisible, instant, and cryptographically sound.
Open any Authored-enabled editor and write as you always do. No interruptions, no special actions. Our system works invisibly in the background.
Every session produces a unique behavioral signature derived from dozens of proprietary biometric features — behavioral signals unique to each individual that no AI can replicate.
Receive a cryptographic certificate with a unique public ID, instantly verifiable by any institution, publisher, or employer.
Authored analyzes dozens of behavioral biometric features — a multidimensional portrait of how your mind controls your fingers.
Captures the micro-timing unique to your neuromuscular system — as personal as a fingerprint, impossible to replicate.
Decodes the pauses, corrections, and rhythm that reveal genuine cognitive engagement vs statistical generation.
Distinguishes the beautiful imperfection of human writing from the statistical uniformity of AI output.
A multidimensional behavioral portrait built from dozens of proprietary signals — impossible to reverse-engineer or replicate.
The fundamental principle
Every human writing session leaves an invisible trace — the neural rhythm of thought-to-finger. This pattern is as unique as DNA and impossible to fake, because it emerges from your neuromuscular system, not from words. Authored captures this trace without ever reading a single word you write.
Live demo
Type in the editor below and Authored will verify you're human. Try pasting text — or hit Simulate AI — to see what happens.
Analysis result
Waiting for input
Type or paste text to see
the analysis result.
How it works: Authored measures the rhythm between keystrokes. Machines type at a perfectly constant pace — humans never do.
This is a concept demonstration. The production Authored platform uses a fine-tuned AI model trained on dozens of proprietary biometric features for forensic-grade accuracy (>99%).
Use cases
Authored integrates where content is created — protecting the authenticity of human work across every industry that values it.
Academic integrity
Give students cryptographic proof of genuine authorship. Protect institutions from false accusations while upholding academic integrity standards — without penalizing ESL students.
Hiring authenticity
Verify that cover letters, assessments, and written submissions reflect genuine candidate capability — not AI-generated proxies that mask the real person.
Content certification
Issue verifiable human-authored badges to articles and editorial content. Unlock the price premium that certified human content commands vs AI-generated articles.
Document provenance
Create an auditable chain of authorship for contracts, declarations, and regulated documents. Establish forensic-grade provenance that stands in court.
Privacy architecture
Authored's zero-knowledge architecture transmits only behavioral metadata — keystroke timing, pause patterns, rhythm signatures. The content you write is never captured, stored, or transmitted.
The content you write is never transmitted, stored, or processed by our servers. Only behavioral timing metadata leaves your device.
Privacy compliance is an architectural constraint, not a policy add-on. We cannot capture what we never receive.
We measure when keys are pressed and released — never which key. Your words are invisible to Authored.
Behavioral signatures are processed in real time and converted to a certificate hash. Raw data is never persisted.
Authored is architecturally incapable of capturing your content. The zero-knowledge design is not a promise — it's a technical constraint baked into the system itself. Compliance-ready by default for GDPR, CCPA, and FERPA.
FAQ
Everything you need to know about Authored, behavioral biometrics, and how we verify human authorship.
Authored is a behavioral biometric certification platform that verifies human authorship in the AI era. By analyzing keystroke dynamics, typing rhythm, and cognitive patterns during writing, Authored issues a cryptographically verifiable certificate confirming that writing was genuinely created by a human — not generated by AI.
Authored analyzes dozens of proprietary behavioral biometric features captured passively during the writing process — including keystroke timing intervals, typing rhythm, pause patterns, correction behavior, and neuromuscular coordination. These signals form a unique fingerprint of human cognitive engagement that AI language models cannot replicate.
No. Authored uses a zero-knowledge architecture — only behavioral metadata (timing signals) is analyzed, never the actual text content. Raw biometric data is processed locally and never transmitted to external servers. Authored cannot read a single word you write.
No. AI-generated text is pasted or injected instantaneously, without the natural timing variability produced by human thought-to-finger movement. Even sophisticated AI agents attempting to simulate keystrokes lack the genuine neuromuscular patterns of human writing — making detection reliable at greater than 99% accuracy.
Authored achieves greater than 99% accuracy with zero false positives. The system is carefully calibrated so that genuine human writers are never incorrectly flagged as AI-generated — protecting authentic authors from false accusations.
Authored serves four primary sectors: Education (verifying student assignment authenticity), Human Resources (validating candidate written assessments), Publishing (certifying journalist and author work), and Legal (authenticating contracts, testimonies, and legal documents).
Traditional AI content detectors analyze what was written — comparing text statistical patterns against AI outputs. This approach produces false positives and is defeated by paraphrasing tools. Authored analyzes how the text was written: a behavioral dimension that AI cannot fake. The result is language-agnostic detection with zero false positives.
Yes. Authored is built on a privacy-first, zero-knowledge architecture. No text content is ever read or stored. Behavioral timing signals are processed with explicit user consent and handled in full accordance with GDPR. The platform does not build persistent user profiles or share data with third parties.
Still have questions? Contact us — we reply within 24 hours.
Authored is in final development. Join the waitlist today and be first in line when we open access — plus get early adopter pricing locked in forever.
No credit card required. No spam. Unsubscribe anytime.
2
Co-founders
67
Biometric features
>99%
Target accuracy
0
False positives