Angèle Angèle
EN FR

Your AI. Your rules.

Angèle lets you define where data is processed, which models may be used, who accesses what and within which framework.

The essentials

Three decisions that are yours.

Where your data is processed

On your infrastructure, in our Private Cloud in Luxembourg, or at a public provider you have authorised for a specific use. That choice is made flow by flow, and it is recorded.

Who accesses what

Access is segmented user by user and aligned with the rights already in force in your organisation. An assistant consults the same documents as the person using it, and only those.

What record remains

Every use, every access and every flow can be logged according to your settings. Enough to answer an internal control or an auditor.

Where it is steered

One administration console, four levers.

Those three decisions do not stay on paper: they are configured, changed and checked from the platform's administration.

  • Users and groupsRights by role and by team
  • Assistants and modelsWhat is enabled, and for whom
  • Data and sourcesSource by source
  • Logs and indicatorsWho did what, with which data
Choosing the model

Local or public, according to how sensitive the data is.

The right model depends on context. Angèle lets you decide use by use, and you can prohibit public models entirely.

Local models

On your infrastructure or in our Private Cloud

Open-source models compatible with the vLLM ecosystem. Your information is not sent to the model's publisher.

  • Processing stays within the defined perimeter
  • No transmission to a third party
  • Suited to sensitive data

Public models

Optional, accessible via API

State-of-the-art models for advanced or highly specialised uses, enabled under your rules.

  • Reserved for non-sensitive tasks
  • Provider conditions verified before activation
  • Framed and reversible activation

The level of control over data varies with the model and the contract. Some public services may retain data or be subject to non-European legislation. We check those conditions with you before any activation.

Traceability

Governed AI must be able to explain itself.

Each step of a use can be logged according to your settings, for internal control as much as for audit.

  1. UserWho?Identity, role, access rights.
  2. RequestWhat?Content, context, intent.
  3. SourcesWith which data?Sources consulted, permissions.
  4. ModelHow?Model used, parameters, tools.
  5. AnswerResultGenerated answer, actions, exports.

Restricted access to logs, retention according to your policies, export available for audits.

Responsibilities

Who does what.

A division written down before the project starts, so nobody discovers their perimeter along the way.

Your organisation

You remain in charge of your usage.

  • The use case and its level of risk
  • The data and the rights to use it
  • Your users' authorisations
  • The security of your workstations and network

Angèle and its partners

We secure the platform.

  • Hosting and its operationPrivate Cloud operated in Luxembourg by Rcube, a support PSF.
  • Orchestration of models and sources
  • Governance and traceability mechanisms
  • Updates and supervision
  • Cybersecurity expertiseRisk analysis and CISO support from Rsecure where your context requires it.

Public model providers

When you enable one.

  • Providing and operating the model
  • Their terms of use and retention
  • The compliance of their own services
Our evidence

Trust is not declared, it is proven.

Hosting in Luxembourg

Private Cloud hosted and operated by Rcube, a support PSF.

ISO 27001 infrastructure

Angèle relies on hosting infrastructure certified to ISO 27001.

Protected data

Encryption in transit and at rest, environment segmentation, retention policies.

Continuous supervision

High availability, backups, anomaly detection and incident management.

Compliance never depends on a platform alone: it also depends on your use case, your data and your organisation. Angèle provides the governance, control and traceability mechanisms that contribute to meeting GDPR and AI Act requirements.

Security questions, in detail in the FAQ

Let's talk

Your AI project starts here.

Your teams, your data, your processes: let's look at how our sovereign AI can make your people's work simpler and move your projects along faster.

Talk to our experts
  • Thirty minutes to qualify the need, the available data and the right starting point.
  • A single point of contact and a multidisciplinary team alongside you.
  • Concrete recommendations, even if the project does not start straight away.