IMBox AI

You build your organization on AI. Build it on IMBox AI.

Built for law enforcement, defense, public administration and any organization with critical security requirements. Run on your hardware, inside your perimeter.

Sovereignty by design

Your AI runs where your data lives

IMBox AI is the only military-grade AI platform built on the assumption that your data never leaves the building. Inference hardware, models, orchestration and applications are deployed inside your security perimeter — and stay there.

100% on-premise

BrainBox is installed in the customer's data center and runs without third-party endpoints.

Zero exfiltration

No telemetry, no anonymized samples and no external training on customer information.

Air-gapped environment

Editors, viewers, players and sandboxing live inside IMBox to remove external dependencies.

Institutional scale

Built for public administration, security and defense, with traceability and operational control from day one.

BrainBox

BrainBox brings AI into your own data center

BrainBox is the inference node for IMBox AI: an autonomous compute unit for running quantized open source models inside your infrastructure, with full control over data, performance, and scaling.

BrainBox cluster

Linear capacity, standard rack deployment

GPUs

128 GB

VRAM

1.500

TFLOPS

32C

CPU

192 GB

RAM

N+1

modular scaling

GPU BOUNCER

The software that turns your hardware into a production AI system

Without it, it's just raw hardware. With GPUBouncer, everything runs from day one.

ModelsAPIOrchestrationScalingSecurityHigh availability

Models

API

Orchestration

Scaling

Security

High availability

The platform integrates everything needed to consume AI in production from day one without taking on the technical complexity of maintenance and scaling: optimized models, API, orchestration, security, monitoring and high availability.

GPUBouncer is the layer that makes this infrastructure work as a single service. It routes each request to the right model, balances load across available instances, preserves sessions to improve latency, monitors model health and applies retries or fallback transparently. When new compute capacity is added to the cluster, the platform brings it into the service so it can keep growing without redesigning integrations.

API

/v1/chat/completionsActive

/v1/embeddingsActive

Models

T2TL

I2TL

A2T

OCR

EMB

Orchestration

Active models3/5

Running inferences24

Load balancingAutomatic

GPUBouncer

Capacity

68%

Performance

2.4k t/s

Scaling

Clusters online10/10

Available capacity75%

Automatic scalingActive

Security

Access levelsEnabled

DataOn-premise

AuditTraceable

High availability

Active instances10/10

Health checksOperational

IncidentsRecovered

One API

An OpenAI-compatible entry point for connecting current or new applications without redesigning integrations.

Plug and play

Add new BrainBox units and capacity grows modularly, with automatic resource detection and enrollment.

High availability

GPUBouncer monitors model health, isolates failing instances and redirects requests transparently.

Smart KV cache

Session-aware routing reuses cache across requests. Lower latency, lower energy.

Category routing

Seven model families routed by purpose: T2TL, T2TF, I2T, A2T, EMB, OCR.

Model Foundry

Open-source models, quantized and tuned in house

Our engineering team maintains a curated catalogue of seven open-source model families, quantized and evaluated specifically for BrainBox. You get the openness of the community and the operational discipline of a vendor.

CodePurposeGPUsVRAM

T2TLText to text, large reasoning4128 GB

I2TLImage to text, large264 GB

A2TAudio to text132 GB

OCROptical character recognition132 GB

EMBEmbeddings132 GB

T2TFText to text, fast132 GB

I2TFImage to text, fast132 GB

Proprietary quantization

Custom pipelines preserve accuracy and shrink models to BrainBox profiles.

Continuous evaluation

Every release is benchmarked on institutional workloads before production.

Managed update cycle

Llama, Mistral and other open-source families integrated and deployed under SLA.

Auditable open-source baseline

All base models are open. No black boxes in the inference path.

The control panel for your sovereign AI

Administration console to govern all of IMBox AI: model catalogue, published assistants, ontologies, permissions, evals and consumption.

Connect AI to your internal services. Safely

Integration layer for querying databases, MCPs, private APIs and document repositories without exposing credentials or moving data outside the perimeter.

OpenAI-compatible. Sovereign by default

Proprietary endpoint compatible with the OpenAI API. Integrate IMBox AI in third-party apps without rewriting code, keeping inference inside the perimeter.

Assistants that remember what matters

Long-term memory layer that allows assistants to remember context, preferences and operational knowledge across time — inside the secure perimeter.

Agent dashboard

Assistant	DocsDocuments	Purpose
CPPublic procurement	84	Interprets tenders, contract files and award criteria with official sources.
NIInternal regulations	126	Answers questions about the organisation's rules, protocols and circulars.
EIInstitutional case files	58	Searches verified files, resolutions and administrative precedents.
JAApplied case law	32	Connects concrete cases with doctrine, judgments and current regulation.
AOOperational analysis	47	Summarizes incident data, internal reports and operational criteria for field teams.

AI productivity for users

IMBox Agentic Assistant

A general-purpose agent ready to work from day one

IMBox Agentic Assistant ships with the core capabilities needed to operate across the IMBox ecosystem: reading documents, querying information, working with images, processing audio, running OCR, generating content and coordinating actions between modules. It does not start as an empty shell; it acts as a transversal agentic layer that helps teams complete real tasks inside the platform.

Agent capabilities

Image viewer

Doc creator

OCR viewer

Audio processor

IMBox management

Maps

+30 more tools

IMBox Recorder & Transcriber

Record, transcribe and summarize every conversation that matters

Native recorder and transcription flow for meetings, interviews, witness statements, emergency calls and operational debriefs. Capture the audio, identify speakers and generate summaries automatically.

Recording and transcription

REC

00:14:32

00:03Speaker 1Can you confirm where the call started?

00:18Speaker 2At the north gate, around 08:40.

01:04Speaker 1Good. Add that to the incident timeline.

IMBox Translator

Translate, summarize, rewrite. In any tone

Upload a document and IMBox translates it, summarizes it or rewrites it. Choose the tone (enthusiastic, friendly, confident) and the writing style (academic, professional, plain) that fits the audience.

English

The report confirms that the perimeter was closed before the second patrol arrived.

All evidence must remain available for review by the command unit.

Spanish

El informe confirma que el perímetro quedó cerrado antes de la llegada de la segunda patrulla.

Todas las evidencias deben permanecer disponibles para su revisión por la unidad de mando.

IMBox Doc Generator

Word, Excel and PowerPoint, written by your AI

Generate office documents in conversational mode. Ask in natural language and IMBox drafts, edits and iterates on the document for you.

Generate procurement brief, budget table and executive deck from Drive context.

DOCX

Brief

XLSX

Budget

PPTX

Deck

IMBox OCR Viewer

Read every document. Even the scanned ones

Native viewer with built-in OCR for PDFs, images and operational paperwork. Indexes, searches and extracts structured information — locally.

Incident Report

Ref. 2026/A-3421

Official

Date

14/05/2026

Unit

Command A3

The report confirms that the perimeter was closed before the second patrol arrived.

All evidence remains available for review by the command unit.

Item	Status
Gate	Closed
Files	Traced
Review	Open

Signature

Operations desk

# Extracted document

## Fields

title: Incident Report

exp: 2026/A-3421

date: 14/05/2026

unit: Command A3

## Summary

Perimeter closed before second patrol arrived.

Evidence available for command review.

## Table

| item | status |

| Gate | Closed |

| Files | Traced |

| Review | Open |

## Signature

Operations desk

IMBox Ontology

Turn fragmented data into a queryable knowledge graph

Ontology builder that ingests data from multiple sources, models it as a knowledge graph and lets you query it in natural language.

IMBox DataBridge

Connect AI to your internal services. Safely

Integration layer for querying databases, MCPs, private APIs and document repositories without exposing credentials or moving data outside the perimeter.

MCP

API

Files

CLI

const db = databridge.database("records");

const caseFile = await db.findOne({
  table: "expedientes",
  where: { ref: "2026/A-3421" },
  audit: true,
});

return assistant.answer(
  "Resume el estado del expediente",
  { context: caseFile },
);

IMBox API

OpenAI-compatible. Sovereign by default

Proprietary endpoint compatible with the OpenAI API. Integrate IMBox AI in third-party apps without rewriting code, keeping inference inside the perimeter.

curl -X POST \
  "https://ai.imbox.local/assistant/api" \
  -H "Content-Type: application/json" \
  -d '{
  "command": "RESTAskAssistant",
  "api_key": "imb_***_4f2",
  "messages": [
    {
      "role": "user",
      "content": "Resume el expediente 2026/A-3421"
    }
  ]
}'

IMBox AI

Bring sovereign AI inside your perimeter

A private intelligence layer for organizations that cannot afford to lose control over their data.