IOanyT Innovations
AWS AMI AI / ML v1.0.0

IOanyT SecureAI

CIS-hardened, GPU-accelerated generative-AI stack that boots and self-configures — zero manual setup.

Delivery method

AWS AMI

Operating system

Ubuntu 22.04 LTS

Architecture

x86_64

GPU

NVIDIA T4 — driver 535 / CUDA 12.2 (validated on g4dn)

Released

July 14, 2026

CIS Benchmark Level 1 hardened — 421 controls applied
NVIDIA GPU ready — driver 535 / CUDA 12.2
vLLM OpenAI-compatible API — default model Qwen2.5-3B-Instruct
Qdrant vector database for retrieval-augmented generation
TLS everywhere via Caddy on port 443
Per-instance API keys + TLS cert generated automatically on first boot

What it is

IOanyT SecureAI is a CIS Benchmark Level 1-hardened, GPU-accelerated Ubuntu 22.04 LTS image with a complete, production-ready generative-AI stack that boots and self-configures with no manual setup. It combines enterprise security hardening with a turnkey LLM inference + RAG environment, so you go from launch to a secured, working AI API in minutes rather than days.

The full stack auto-starts on boot via systemd. On first launch each instance generates its own unique API keys and TLS certificate — there are no shared or hardcoded credentials baked into the image.

SecureAI is published by IOanyT Innovations, a company certified to ISO/IEC 27001:2022 (Information Security Management) and ISO 9001:2015 (Quality Management) — both issued by QRO and valid to August 2029. The same certified processes that govern our client delivery govern how this image is built, hardened, and maintained.

Who it's for

Teams that need a private, secured LLM endpoint without assembling and hardening the stack themselves.

Regulated / security-conscious environments that require CIS-benchmarked baselines and audit evidence out of the box.

Builders of RAG applications who want an inference API and a vector database pre-wired behind TLS.

Teams that want OpenAI-compatible inference on their own AWS account — data stays inside your VPC.

Why it's different

Secure by default

CIS Level 1 (421 controls), key-only SSH, auditd enabled, no hardcoded credentials, and per-instance secrets.

Turnkey, not a toolkit

vLLM, Qdrant, and Caddy are pre-integrated and auto-start. No wiring, no reverse-proxy config, no cert wrangling.

OpenAI-compatible

Point any existing OpenAI client at your instance's /v1 endpoint and it just works.

Evidence included

Ships with CIS conformance evidence and an SBOM for your compliance and supply-chain reviews.

ISO-certified publisher

SecureAI is built and maintained by IOanyT Innovations, an ISO/IEC 27001:2022-certified company — information security governed by an independently audited ISMS, with ISO 9001:2015-certified quality management behind every release (QRO, valid to Aug 2029).

Security & compliance

CIS Benchmark Level 1 hardening applied to Ubuntu 22.04 LTS (421 controls).

Key-only SSH — password authentication disabled.

auditd enabled for system audit logging.

No hardcoded credentials — VLLM_API_KEY and QDRANT_API_KEY generated uniquely per instance on first boot.

TLS everywhere — all application traffic terminates at Caddy on port 443 with an auto-generated certificate.

Ships with CIS conformance evidence and a Software Bill of Materials (SBOM).

Published by an ISO/IEC 27001:2022-certified company — IOanyT Innovations' information security management system is independently audited (QRO, valid to Aug 2029).

You remain responsible for your AWS account security (IAM, security groups, network exposure, patching cadence after launch) under the AWS shared-responsibility model.

What's included

Software Version
Ubuntu (CIS Benchmark Level 1 hardened — 421 controls) 22.04 LTS
NVIDIA GPU driver 535
CUDA 12.2
vLLM (OpenAI-compatible inference API)
Qwen2.5-3B-Instruct (default model)
Qdrant (RAG vector database)
Caddy (reverse proxy, automatic TLS)
Docker Compose (stack orchestration)
auditd (audit logging)

Recommended instances

g4dn.xlargeg4dn.2xlarge

Ports

  • 443/tcp — HTTPS — LLM API, Qdrant, all app traffic (Caddy reverse proxy, TLS)
  • 22/tcp — SSH — key-only administrative access

Pricing

Pay-as-you-go

Hourly (pay-as-you-go)

Software billed hourly via AWS Marketplace, on top of standard EC2/GPU instance costs.

Frequently asked questions

How is IOanyT SecureAI delivered on AWS Marketplace?

IOanyT SecureAI is delivered as an Amazon Machine Image (AMI). You launch it as an EC2 instance directly from the AWS Marketplace listing.

How much does IOanyT SecureAI cost?

Pay-as-you-go — Software billed hourly via AWS Marketplace, on top of standard EC2/GPU instance costs. Software charges are billed through AWS Marketplace; you also pay standard AWS infrastructure costs.

How do I access IOanyT SecureAI after deploying?

Connect over SSH as "ubuntu" using your EC2 key pair. See the deployment guide for full steps.

What operating system does IOanyT SecureAI run on?

Ubuntu 22.04 LTS (version 1.0.0).

What support is available for IOanyT SecureAI?

Business-hours email support. Email support@ioanyt.com for help.

Support

Business-hours email support

Questions or issues? Email support@ioanyt.com.

Ready to deploy?

Launch it in your own AWS account, or talk to our team about a custom hardened image for your stack.

Request early access

Runs entirely inside your AWS account