The only proof of human that still works
Every existing proof of humanness is defeated by AI. FaceSign is the one that isn't.
AI broke every proof of human.
The tools built to prove humanness — CAPTCHAs, document checks, selfie matching — are now defeated by the same technology they were designed to stop.
CAPTCHAs
Solved faster by bots than by people. AI solves CAPTCHAs more accurately than humans. The original proof of human is now a proof of nothing — a speed bump that slows real users and does nothing to stop automated attacks.
Choose the chairs that look the comfiest, but actually aren't
Document Verification
Synthetic IDs pass at scale. AI-generated documents defeat automated checks. Static document verification is a single point of failure — once forged, no downstream system catches it.
Selfie Matching
Deepfakes fool liveness checks. Face-swap and video injection attacks defeat liveness checks that rely on a single frame or a head turn. The camera sees a face, but it's not a person.
AI Agents
No human in the loop. Autonomous agents execute transactions, change accounts, and sign documents. No existing verification checks whether a human authorized the action — because none was designed for a world where the requester might not be human.
The Problem
AI broke every proof of human
The tools built to prove humanness — CAPTCHAs, document checks, selfie matching — are now defeated by the same technology they were designed to stop.
- AI solves CAPTCHAs faster than humans — the original proof is obsolete
- Synthetic IDs pass automated document verification at scale
- Face-swap attacks defeat single-frame liveness checks
- AI agents execute sensitive actions with no human in the loop
The Solution
A live conversation can't be faked
FaceSign proves personhood through an unpredictable AI conversation — analyzing 30+ signals that can't be pre-rendered, injected, or automated.
- Unpredictable live dialogue defeats every pre-rendered attack
- Frame-by-frame temporal analysis catches video injection in real time
- Micro-expression timing proves physiological presence — not a photo or mask
- Behavioral signals confirm the person is acting freely, not under coercion
From action trigger to cryptographic proof in 30 seconds
FaceSign provides the trust layer for sensitive actions. When human authorization is required, verification is instant and unfakeable.
Authorization Required
A sensitive action — transaction, account change, legal signing — triggers the human verification gate.
Conversational Proof
User enters a 30-second unpredictable, real-time dialogue. No pre-rendered attack can pass.
Humanity Confirmed
Six AI models confirm: live human, real identity, no deepfake, no coercion, acting freely.
Attestation Issued
Cryptographic proof of human authorization returned. Tamper-evident, immutable, and auditable.
See the proof-of-human primitive in action.
By the Numbers
Verification that scales
Average verification time
Fraud signals per session
Always available
Signals analyzed per session
Conversational Liveness
Unpredictable real-time dialogue that no deepfake, replay attack, or bot can pass. The challenge changes every time.
Temporal Analysis
Frame-by-frame video analysis proves physiological presence — catching injection attacks that single-frame checks miss.
Intent Verification
Behavioral signals confirm the person authorized the action willingly — detecting coercion, coaching, and duress in real time.
Cryptographic Attestation
Tamper-proof proof that a specific human authorized a specific action at a specific time. Immutable and audit-ready.
AI Agent Gateway
The human-in-the-loop primitive for autonomous AI agents. Verify authorization before the agent executes.
Anti-Automation Layer
Defeats CAPTCHA solvers, document forgers, and selfie injectors — the entire AI-powered attack surface.
Zero-Install Deployment
Browser-based, no app download. Single API call. Works across any platform, device, or workflow.
Benefits
The trust layer for the AI era
30 Seconds, Not 30 Minutes
Verification completes before the user loses patience. No queues, no hold music, no manual review.
Defeats What CAPTCHAs Can't
A live conversation can't be solved by a bot, injected with a deepfake, or bypassed with a stolen document.
Replaces Manual Review
Automates the judgment call your team makes manually — at scale, around the clock. No per-case agent cost.
Fewer False Blocks
Legitimate users verify in seconds instead of getting locked out. Higher completion rates on sensitive actions.
Proves Intent, Not Just Identity
The only check that confirms the person is acting freely — catching coercion that every other tool misses.
Evidence-Grade Compliance
Full audit trail with tamper-proof session recordings. Meets PSD3, eIDAS 2.0, and EU AI Act requirements.
Built for critical moments
Explore more use cases
Every proof of human is broken except this one
See FaceSign verify a real, willing human in 30 seconds.