What is CLAVI?

CLAVI is a Swiss hardware and software company developing a Personal Vault for digital life. Its public documentation describes local-first hardware, protected key workflows and proprietary CLAVI AI intended for businesses and high-net-worth individuals. The Clavi app is presented as an on-device, multi-agent system extending that local-first objective into everyday AI workflows. These are product design statements; production behaviour and security assurance depend on the deployed version, configuration and supporting engineering evidence.

1. What CLAVI is

CLAVI’s public materials describe a Swiss company developing sovereign-grade hardware and software for CLAVI devices and supported user devices. The documented product surface includes:

• Edge encryption hardware. Devices intended for on-premises key generation, signing and custody. CLAVI describes air-gapped workflows, biometric authorization and a post-quantum security objective; the algorithms, implementation and assurance scope require version-specific technical documentation.
• ClavOS. A zero-knowledge operating system that runs on CLAVI’s hardware. CLAVI’s stated design objective is to keep the company outside the user’s key and protected local-content paths after delivery; this is distinct from the limited commerce and support records an operating company may hold.
• The Clavi app. A multi-agent application intended to run protected workloads locally on supported computers and smartphones. Any cloud, telemetry, update or optional-service boundary should be stated for the specific production version.

CLAVI positions the product for businesses seeking on-device AI for sensitive workflows and high-net-worth individuals — including family offices — seeking greater physical control over digital assets and private information. The company operates from Switzerland, where Article 13 of the Swiss Federal Constitution protects privacy [2], outside the European Union and the Five Eyes intelligence-sharing alliance. CLAVI presents that Swiss legal setting as one layer of its architecture; it is not a product certification or immunity from applicable law. For a fuller account of the company’s origins, see About CLAVI [1].

In one line: CLAVI is designed as a Personal Vault for users who want protected keys, private data and proprietary CLAVI AI workloads to remain on hardware they control.

2. Local-first security

The phrase “local-first” is the design center of the company. In CLAVI’s public model, protected key operations, AI inference, document handling and persistent context are intended to run on user-controlled hardware. It should not be read as proof that every feature works without connectivity or that no update, support, commerce or optional-service data ever reaches a third party.

Local-first is not the same thing as “offline-capable.” CLAVI’s stated objective is to make supported critical workflows operable without a live network connection, while treating connectivity as an explicit, bounded dependency when a feature requires it. Exact offline coverage needs testing against the production release.

This matters because the threat model for serious users has expanded. The risk is no longer just a compromised browser. It is a compromised supply chain, a coerced cloud provider, a subpoenaed log, a leaking model API, or a future cryptographic break against keys that were generated on a connected machine. A local-first architecture can remove categories of exposure by removing the assumptions they depend on. It does not erase the separate order, invoice, support, or regulatory records an operating company may need. See Zero-Knowledge Architecture, Data Residency, and CLAVI’s data-minimization analysis for the underlying boundaries.

3. Edge encryption hardware

CLAVI’s hardware is designed around three architectural choices.

Post-quantum security objective. CLAVI describes key-generation choices intended to resist classical and future quantum adversaries. “Quantum-resistant” is not a timeless guarantee: assurance depends on the named algorithms, parameters, implementation, migration plan and current cryptanalytic evidence. CLAVI’s stated design keeps key generation on the designated device.

Air-gapped operation. CLAVI describes critical key operations as isolated from network-attached systems. A correctly implemented air gap can reduce remote attack surface; it does not remove firmware, supply-chain, removable-media, configuration, side-channel or physical risks. See Air Gap and Hardware Custody for the canonical definitions.

Biometric authentication. CLAVI describes biometric approval as one authorization control for sensitive operations. Its effectiveness depends on enrollment, matching thresholds, liveness controls, fallback and recovery paths; possession of a stolen device remains a security event rather than proof that the device is inert.

These choices are intended to work as layers. No single layer establishes the security of the whole system: cryptographic choice, isolation, authentication, recovery, updates and operational configuration all require evidence and testing.

4. The Clavi app and local AI stack

The Clavi app is CLAVI’s software product. CLAVI describes it as a local multi-agent AI stack for supported user devices. Platform compatibility, hardware requirements and setup behaviour should be taken from the current release documentation rather than inferred from a device’s year.

The same local-first objective governs the app. CLAVI says model execution, orchestration, memory and document context are intended to remain local for supported workflows. A publishable assurance claim needs version-specific network testing and documentation of updates, diagnostics, optional integrations and telemetry defaults.

That boundary can matter in regulated and high-trust contexts because a conventional hosted assistant processes prompts on provider-controlled infrastructure. Whether a particular Clavi workflow remains entirely local must be verified for that workflow and release.

5. Airplane-mode operation and sandboxing

The Clavi app is intended to support offline work, including selected airplane-mode workflows. Document review, analysis, research and drafting may be suitable local tasks, but feature availability, model assets, licenses, updates and external data dependencies should be tested before relying on offline continuity.

CLAVI documentation describes sandboxed workload environments intended to isolate agent execution from the rest of the device. Sandboxing can limit some consequences of malicious content or compromised tools, but it does not by itself prevent prompt injection, data exfiltration or local privilege escalation. The actual boundary depends on operating-system controls and their implementation.

6. Multi-agent orchestration

CLAVI describes the Clavi app as a coordinated multi-agent design rather than a single monolithic model; the deployed design requires version-specific verification.

CLAVI describes a master agent with persistent memory around which task-specific models can be coordinated. The examples include document extraction, reconciliation, summarisation and code generation; their availability and quality are product claims that require release-specific evaluation.

CLAVI describes an orchestration layer that routes work among the master agent and specialist models according to task and device constraints. The routing logic, privacy boundary and performance are engineering characteristics to document and test, not assumptions established by the “multi-agent” label.

7. Proprietary RAG tagger and persistent context

The orchestration layer would not be useful without large, reliable context. Standard retrieval-augmented generation (RAG) systems break down quickly on real corpora: they retrieve the wrong passages, lose track of relationships between documents, and produce shallow answers when the question requires reasoning across an entire archive.

CLAVI says it uses a proprietary RAG tagger to organize heterogeneous corpora such as contracts, emails, financial statements and codebases into a queryable substrate. Claims about its research history, retrieval quality, context scale and competitive performance require dated documentation and reproducible benchmarks. The same stated approach underlies the proprietary CLAVI AI’s local knowledge-management design.

This is the layer where local AI starts to become useful for serious professional work, rather than a parlour trick.

8. Real-world use cases

The combination of local-first execution, multi-agent orchestration, and long persistent context maps onto several professional workflows that have historically been hard to serve well from the cloud.

• Accountants reviewing a client tax history on a supported local device, subject to the workflow’s documented storage and network boundaries.
• Law firms conducting local document review where professional-secrecy, privilege, access-control and device-management requirements have been assessed independently.
• Family offices running internal analysis on holdings, beneficiaries and inheritance structures, with the intent of reducing external-assistant exposure subject to verified data flows.
• Professional services teams — including some of CLAVI’s own back-office functions — deploying the same software internally to handle their own sensitive records.

These are illustrative scenarios, not compliance or security assurances. Each deployment still needs access controls, endpoint security, backups, legal analysis and verification of the actual data flow.

9. Why CLAVI is difficult to replicate

CLAVI presents its integrated stack—ClavOS, CLAVI hardware and a proprietary RAG tagger—as a product differentiator.

Integration across hardware, operating-system and retrieval layers can be difficult. Whether CLAVI’s implementation creates a durable advantage is a market and engineering question requiring comparative evidence.

The relevant evaluation is therefore the integrated outcome: version-specific security evidence, retrieval quality, supported context, recovery behaviour and operational usability—not the number of proprietary layers alone.

For the comparative-architecture argument relative to mass-market hardware wallets, see Why CLAVI Isn’t Competing with Ledger. For the buyer’s decision framework, see Should You Buy a Clavi? Sovereignty Buyer’s Guide 2026. For the jurisdictional layer, see Jurisdiction as a Service.

If your use case touches any of this — local AI, sensitive data, digital wealth, or all three — the next step is to contact the team directly.

Frequently Asked Questions

Q: What is CLAVI?
A: CLAVI is a Swiss hardware and software company developing a Personal Vault for digital life. Its documentation describes local-first key workflows, protected data and proprietary CLAVI AI for businesses and high-net-worth individuals. Production capabilities and assurance depend on the deployed version, configuration and engineering evidence.

Q: What does the Clavi app do?
A: CLAVI documentation describes an on-device, multi-agent application with local orchestration, persistent context and task-specific models. Exact platform compatibility, offline feature coverage and performance require version-specific product documentation and testing.

Q: Is CLAVI cloud-based?
A: CLAVI is local-first by design. Its stated objective is to keep protected key operations and proprietary CLAVI AI inference on user-controlled devices. That does not prove that every feature is cloud-independent or that no business, support, update or optional-service data is ever transmitted; those boundaries require version-specific documentation.

Q: Does CLAVI work offline?
A: Offline operation is a documented design objective for supported local workflows. Which functions work in airplane mode, and which need updates, synchronization or external services, should be verified for the production version.

Q: Who is CLAVI for?
A: CLAVI positions its products for businesses and high-net-worth individuals with serious privacy, compliance and continuity requirements, including family offices, regulated professionals and organisations seeking on-device AI. Suitability depends on the deployed version, documented controls and the user’s own risk assessment.

Q: What makes CLAVI difficult to replicate?
A: CLAVI describes an integrated stack combining its hardware, ClavOS and a proprietary RAG tagger. The maturity, security and comparative advantage of that stack require benchmarks, architecture documentation and version-specific independent testing.