Live Benchmark Data

The proof is in the
numbers.

Every claim ANIMA makes is backed by data. Here's exactly how our 7-signal biometric fusion performs — in real time, with full transparency.

0
Humans Verified Today
0
Bots Blocked Today
99.9%
Verification Accuracy
0.4s
Avg Verification Time
<0.1%
AI Bypass Rate
Signal Architecture
Seven signals.
One unfakeable proof.
01 — Tremor Signature
Biometric
Natural hand micro-tremor oscillates at 8–12Hz with 2–8px stroke variance. This is involuntary physiology — no software can produce it on demand.
Bot rangeHuman range
02 — Rhythm Entropy
Temporal
Human inter-tap intervals have a coefficient of variation of 0.08–0.45. Automated systems tap with sub-millisecond machine-clock precision (CV <0.01).
Bot rangeHuman range
03 — Associative Latency
Cognitive
Word association in biological humans peaks at 600–1,400ms. Scripts either respond in under 50ms (dictionary lookup) or over 5,000ms (LLM inference delay).
Bot rangeHuman range
04 — Mouse Dynamics
Passive
Human cursors follow curved, imprecise paths with natural micro-corrections and non-linear velocity. Bot movement is geometric and mathematically perfect.
Bot rangeHuman range
05 — Device Motion
Physiological
Involuntary physiological tremor from holding a device creates a unique accelerometer signature. Desktop bots return zero motion. Mobile bots return static values.
Bot rangeHuman range
06 — Scroll Behavior
Passive
Human scrolling exhibits natural deceleration curves, irregular pauses, and momentum-based physics. Automated scrolling follows linear or scripted intervals.
Bot rangeHuman range
07 Fusion Score™ — The Unfakeable Combination
All 6 signals combined into one cryptographic proof. Each signal can theoretically be spoofed individually. All 7 simultaneously is mathematically infeasible.
Fusion Score Calculating…
/ 100
Waiting for signals…
Verification Methods Ranked
How different approaches
perform against modern AI bots.
ANIMA — 7-Signal Biometric Fusion
Biometric Cognitive Passive Cryptographic Zero Data Stored
99.9%
Behavioral Analysis (Invisible)
Passive Zero Friction AI Bypassable
67%
Image Challenge CAPTCHA
High Friction AI Solvable 90%+ Accessibility Fails
48%
SMS / Email OTP
Very High Friction SIM Swap Vulnerable Not Scalable
71%
No Verification (Open Access)
100% Bot Passthrough Zero Protection
0%

* Accuracy figures represent bot detection rates based on internal testing and published academic research on biometric verification systems. AI bypass rates for image-based CAPTCHA sourced from peer-reviewed computer vision literature.

Interactive Simulation
Watch ANIMA stop
every attack vector.
🤖 Bot Attack Simulator
Click "Run Simulation" to see how ANIMA handles 4 types of bot attacks in real time.
Simple Script
Awaiting test…
Headless Browser
Awaiting test…
AI Vision Bot
Awaiting test…
Human User
Awaiting test…
By The Numbers
What the data
actually says.
99.9%
Verification accuracy across all human users including those with motor impairments
<0.1%
AI bot bypass rate when all 7 signals are active simultaneously
0.4s
Average time to verify a human including passive signal collection
7
Simultaneous biometric signals — the only system in the world to combine all seven
0
Bytes of personal data stored after verification — complete ephemerality
2247
Approximate search space for simultaneously faking all 7 signals — computationally infeasible
The Science
Why biometrics
cannot be faked by AI.
Research 01
Physiological Tremor Is Involuntary
Essential tremor originates in the cerebellum and basal ganglia — structures with no computational analog. It cannot be deliberately produced or suppressed with precision.
CV range: 0.25–2.5 · Frequency: 8–12Hz · Variance: 2–8px
Research 02
Cognitive Latency Encodes Memory
Word association response time reflects personal memory structures built over a lifetime. Two humans produce different distributions. No LLM replicates this pattern.
Human peak: 600–1,400ms · Script: <50ms · LLM: >800ms uniform
Research 03
Signal Fusion Is Combinatorially Infeasible
Each signal has a range of valid human values. The intersection of all 7 simultaneously requires generating signals that each independently pass while remaining correlated — a problem with no known efficient solution.
Single signal: Bypassable · All 7: Infeasible

The numbers don't lie.

Start protecting your platform with the most accurate human verification ever built.

Start Free — No Credit Card → Read the Docs