Skip to content

AI within the perimeter the company chooses.

VERIAX installs on the customer's infrastructure, in its private cloud, or hosted and managed by THUX CODE.

Talk to the technical team Learn more about Control
Two professionals in a corporate corridor with a server room and facility visible through glass walls.

The problem

Many widely used AI solutions run on public cloud: company documents, contracts and know-how leave the perimeter. For many sectors this is not an acceptable option.

Bringing AI inside the perimeter, on the other hand, seems to require complex infrastructure and specialist skills.

The solution

VERIAX adapts to the existing infrastructure: the same framework, three installation modes. In all of them the customer sets the data perimeter, and the setup is fully managed by THUX CODE.

Zerodata sent to third parties, with local models.

The three modes

On-premise

physical or virtual servers, including Kubernetes
VERIAX runs on the company's servers, physical or virtual, including a Kubernetes cluster already in use. It is the choice for networks with restrictive access rules or when GPUs are needed to run models locally. With local models, data does not leave the company data center.

Customer private cloud

AWS · Azure · GCP, on the customer's account
VERIAX installs in the customer's private cloud. Compute resources remain under the customer's control, with management and billing on its own account. No data on shared public infrastructure.

Hosted by THUX CODE

data center managed by THUX CODE
For those who prefer not to manage infrastructure, VERIAX is hosted in THUX CODE's data center, with EU data residency and no use of public cloud.

One framework. The customer chooses the perimeter.

Control

Data stays with the company.
The company chooses the models.
Key actions are tracked.

VERIAX is not tied to a single model: the model can be changed from the interface, even while the system is in use, choosing among the main cloud providers or open-source models running locally.

Setup timeline

  1. 01

    Infrastructure installation and configuration

    Weeks 1–2

  2. 02

    Configuration of agents, knowledge base, connectors and permissions

    Weeks 3–4

  3. 03

    Field trial of the first use case and collection of indicators

    Weeks 5–6

  4. 04

    Go-live, rollout to users, training

    Weeks 7–12

Indicative timelines: the first use case is typically tested in the field in about 6 weeks and goes into production in about 12.

In brief

VERIAX is a framework for enterprise AI agents that installs within the perimeter chosen by the customer: on-premise on its own infrastructure, including Kubernetes, in its private cloud (AWS, Azure, GCP) or hosted in THUX CODE's data center. In every mode the data perimeter is defined by the customer; with locally run models, no data is transmitted to third parties. The model is chosen and changed from the interface, including open-source models running locally.

One framework. The customer chooses the perimeter.

Talk to the technical team Learn more about Control