What is CheshireCat® and how is it different from facial recognition surveillance?

CheshireCat® is Mad Systems' privacy-forward recognition and context continuity system. It determines that a visitor at one location is the same visitor who was at a previous location — enabling narrative continuity across a visit — without attempting to identify who the visitor is by name or building a persistent identity record. This is architecturally distinct from surveillance facial recognition. Processing is local, data is not transmitted externally in default configuration, and retention is configurable to institutional policy.

How fast is CheshireCat® recognition?

CheshireCat® achieves recognition and context restoration in approximately 0.5 seconds — fast enough to be invisible to the visitor experience. The response time is achieved through local processing without external data transmission.

Does CheshireCat® store biometric data?

CheshireCat® uses local processing with configurable retention settings. In default configuration it does not retain biometric data beyond the visit session. Retention policies are configurable to match institutional governance requirements and visitor trust obligations. The system is designed for privacy-forward operation — it enables personalization without the data collection footprint of conventional identification systems.

What recognition modalities does CheshireCat® support?

CheshireCat® supports optical recognition for session context continuity, combined with RFID, NFC, QR codes, and barcode reading. The full recognition stack builds a useful picture of visitor context — exhibit focus, engagement depth, location history — without assigning a persistent identity to the visitor.

Is CheshireCat® patented?

Yes. Mad Systems holds a US patent covering the CheshireCat® recognition and context continuity methods, specifically the privacy-forward approach to session continuity in public venue environments.

CheshireCat® - Recognition and Context Continuity

CheshireCat® context continuity and privacy-forward recognition for physical venues

CheshireCat^® is the Mad Systems recognition and context continuity layer for physical venues. It helps environments determine what guest context applies, what continuity should be preserved, and what should happen next, without reducing the experience to a generic surveillance model.

Depending on deployment needs and venue policy, CheshireCat^® can work with multiple context signals, including optical recognition where appropriate, RFID and NFC, QR and barcode interactions, beacons, ticket-linked identifiers, and other venue-defined inputs. The goal is not recognition for its own sake. The goal is continuity, accessibility, personalization, safety, and operational responsiveness in a privacy-forward framework.

Privacy architecture

Privacy by Design, Not by Policy

CheshireCat^® is built around a specific principle: recognition in public environments must be designed to minimize risk, not managed after the fact. Default operation is anonymous. Processing is designed to stay on-premises. Biometric data is not retained in default deployments. Consent structures are configurable to venue policy and applicable legal requirements.

Default anonymous operation

Most venue operational decisions do not require identifying a person. Language preference, depth level, accessibility mode, and crowd flow management can all be handled without individual identity. CheshireCat^® operates anonymously by default.

Local-first processing

All recognition processing happens on-premises, on local hardware, with no internet connection required. No biometric data is transmitted to external servers or cloud services. Venues can operate CheshireCat^® on an isolated, air-gapped network.

No persistent biometric storage

In anonymous mode, the system maintains short-lived session context only, preserving accessibility and language preferences across exhibits during a single visit, without building a long-term identity profile. Biometric data is not retained beyond the immediate session in default configuration.

Explicit opt-in for identity

Identity-linked personalization and longer-horizon continuity require explicit visitor opt-in, with a clear explanation of purpose, a defined retention limit, and an accessible deletion process. Covert profiling is not supported.

Data minimization

CheshireCat^® is designed to collect only what is required for the current operational purpose. When that purpose ends, data is not retained. This approach is designed to support privacy-forward deployment and can be configured to align with applicable venue policies and jurisdiction-specific requirements, including frameworks such as GDPR, CCPA, and BIPA where relevant.

Redacted occlusion

Venues can count, route, and manage visitor flow without retaining any identifying image. Anonymous visitors receive the same operational benefit as identified visitors: improved pacing, accessible content, and flow management, without any identity processing.

Capabilities

Recognition and Context Features

~0.5 Second Recognition

Response time of approximately 0.5 seconds, suitable for natural guest interactions at any venue touchpoint.

Local processing

Recognition processing happens on local hardware. In default configuration, no biometric or identification data is transmitted to external services or cloud infrastructure.

No Internet Required

Operates on completely isolated, air-gapped networks. Zero cloud dependency.

Network isolation

Can operate with no external network connections. Suitable for venues with strict data sovereignty, government, heritage, or institutional requirements.

Context Handling

Maintains visitor context, preferences, and consent state within a session, enabling continuity of language, accessibility mode, and personalization across exhibits during a visit.

WorldModel™ Governed

Every recognition event passes through WorldModel™ governance layers. Consent structure and policy rules are enforced at the architectural level.

Deployment context

Designed for Privacy-Forward Public Environments

CheshireCat^® is designed to support privacy-forward deployment in physical venues. It can be configured to align with venue policy, consent structure, operational requirements, and jurisdiction-specific legal obligations. Local processing and policy-bound implementation options allow venues to choose an approach appropriate to the legal context and visitor expectations of each environment.

CheshireCat^® can be implemented in ways that reflect applicable privacy obligations, including consent-aware operation, data minimization, local-only processing, defined retention limits, and audit support through WorldModel™. Implementations are configured per venue and per jurisdiction. Mad Systems does not make blanket representations about compliance with any specific legal framework in all possible deployments. Venues are responsible for configuring and operating the system in accordance with applicable law.