Xeretron AI Server

Resources & commercial kit

Tools to decide with your own numbers.

Documents and materials to take the conversation to your committee. Everything we publish here is tagged by validity level: indicative, under validation, or pending evidence.

Interactive

Start with your organization's numbers.

Quick infrastructure evaluator

Answer eight technical questions and get a preliminary estimate of the node you need: GPU, memory, storage, and offline viability.

Indicative result; final sizing requires evaluation with our team.

Request evaluation

Public documentation

Material for your internal committee.

Commercial one-pager

One-page summary: value proposition, cloud vs private comparison, capabilities, sectors, plans, and process. Ideal to forward internally.

View one-pager

Architecture summary

Public diagram of the node, specialized engines, and security controls for IT review.

View on site

Presales question matrix

The eight decisions that define technical sizing for a private AI deployment.

View the matrix

Cases by sector

Context, problems, typical cases, and honest limits for education, health, government, companies, and labs.

Explore sectors

Partner kit

What the program includes, ideal partner profile, and what we're not yet at this stage.

View program

Claims policy

What we claim, what requires evidence, and what we never publish.

View policy

Transparency

Our claims policy.

A site that promises everything doesn't deserve trust. This is how we classify what we say.

Acceptable

What we claim

  • Designed to run AI on private infrastructure.
  • Integrates models, voice, documents, APIs, and operations.
  • Greater control over models and data.
  • Can improve cost predictability on steady workloads.
  • Local and hybrid operation with guided deployment.
Requires evidence

What we contextualize

  • Savings percentages and uptime.
  • Concurrent users and tokens per second.
  • Install time and impact metrics.
  • Specific regulatory compliance.
  • OpenAI compatibility: supported endpoints only.
Prohibited

What we won't say

  • “100% secure” or “impossible to hack.”
  • “Complies with all laws.”
  • That it replaces doctors, teachers, or officials.
  • That no data can ever leave under any configuration.
  • That any model runs on any hardware.

Need something that isn't published yet?

Ask in the initial conversation. We share what's public and clearly tell you what doesn't exist yet.

Talk to the team