V-PROOF's cryptographic infrastructure
for the AI economy.
Evidence captured at the source, verifiable integrity and third-party verification, built for auditors, regulators and expert witnesses.
From the operational event to verifiable evidence.
Identity, context, and cryptographic fingerprint, organized in a chain of evidence whose integrity can be independently verified.
-
01
Identity and human validation
The action is linked to an owner and an operational context.
Authenticated Identity · Role · Human in the Loop -
02
Governance context
The event is placed within its institutional framework.
Entity · Classification · Policies -
03
Capture in the flow
The cryptographic fingerprint is generated at the point of origin, with data minimization.
API-first · Cryptographic fingerprint · At origin -
04
Structuring the Evidence
The fingerprint and its context are organized into a verifiable package.
Package · Metadata · Chain of Custody -
05
Registration, verification and visible seal
The evidence is linked to a verifiable cryptographic record and the V-Seal® visible on the document.
Cryptographic record · Independent verification · V-Seal®
Inside your architecture. Under your technical control.
Documented interfaces and verifiable events. Each integration defines what data is processed, what remains within the perimeter, and what verifiable cryptographic evidence is recorded.
POST /v1/seal
GET /v1/verify/{id}
Sealing and verification built into your systems. REST, JSON, webhooks, idempotence, and traceability.
client.seal(asset, context)
client.verify(evidence)
A common interface for applications and processes. Typing, error handling, and retries.
vproof seal --reference $BUILD_ID \
--environment production
Evidence at specific points in the cycle: CI/CD, CLI, policy gates. Evidence by design.
ENV test
MODE non-production
Integration testing before working with real data: test environment, simulated data, and separation of environments.
Technical evidence that stands up to external scrutiny.
Cryptographic fingerprints and verifiable chronology, structured so that authorities, auditors and experts can verify the integrity of the evidence without accessing internal systems.
Technical evidence supports the validation of controls before authorities, auditors, and competent bodies. It is designed to be submitted as electronic evidence; it does not replace conformity assessments, the legal basis for processing, or legal advice.
Sixteen modules. A single layer of evidence.
Seal each asset when a V-PROOF flow confirms it: cryptographic fingerprint, timestamp and verifiable evidence.
Compare any asset with its cryptographic fingerprint, without accessing the original.
Assistant powered by the organization's LLM. It filters every message before it is sent and reviews every response, with a human/AI ratio.
Authorship, role, and regulatory context of each asset generated. Compliance starts with the individual.
Analyze changes before branching, evaluate secrets, dependencies, and declared AI use and associates each commit with cryptographic evidence within its context.
MLflow models in the registry: six checks without an LLM, pre-production approval, and V-Seal® for approved versions, along with their files, data, and results.
Audit of agents on six managed platforms, read-only: eight controls without LLM, one approval per configuration, and V-Seal® for the approved configuration. Vertex AI, validated in a live environment; the rest are undergoing validation.
ISO 9001, 14001, 45001, 27001 management systems, or the standard of your choice: step-by-step signatures along the margin of each document, a QR code indicating whether a copy is still valid, and a sealed package for certification.
It allows you to record commits and approved versions with verifiable evidence. Chain of custody for EU CRA and code audits.
The platform's AI workflows follow the same audited call path.
Prompts, outputs, corpora, and development rules, with traceability of usage.
Assets with metadata, timestamps, provenance, and verifiable evidence. Audits in one place.
Documents, teams, and assets organized by project, with version control and traceability from the source.
Every call to a provider passes through a single server module and logs its token usage: thirty days of metrics per system.
API and SDK with all endpoints, GitHub App for repositories, MLflow for models, and discovery from AI providers.
Documented and executable REST API, and native SDKs. The evidence layer where your teams are already working.
V-PROOF Desktop. Evidence from the workplace.
Generate fingerprints, prepare evidence and manage your records from your computer, on macOS, Windows and Linux. Deployment, communications, and access controls are tailored to each organization's policies.
Seal and verify from an installed app, without relying on browser sessions. Filtered chat, a window to seal any file, and a folder that seals everything that lands in it.
Configurable according to the organization's proxy, firewall, TLS, and traffic inspection policies.
The cryptographic fingerprint is calculated on the device. With connectivity, the reference and defined metadata are transmitted.
Distribution, updates, and authentication integrated with SSO, the directory, MDM, and available corporate tools.
Governing AI is not about documenting it.
It's about being able to prove it.
The governance module transforms the evidence that the platform already produces—every trade logged and audited; approvals are recorded and selected evidence can be sealed—into the dossier that regulators, insurers and councils demand.