Live on the Azure Marketplace

Your own private AI.
In your Azure subscription.

A ready-to-run Azure Marketplace appliance: open-weight AI models, a private chat interface, and an honest model catalogue, run entirely from the Azure portal.

Pick a size and deploy the appliance into your own subscription, from a Tesla T4 at pennies per hour up to eight MI300X. The A100 size is benchmarked for gpt-oss-120b at ~140 tokens/sec. Learn more about the appliance.

Microsoft Partner
Available on Microsoft Marketplace
azureuser@caleta-appliance

azureuser@caleta-appliance:~$ llm

--- status -----------------------------

engineUP on 127.0.0.1:8080 (key required) serving: Qwen3-Coder-30B.gguf
restoreoff (use 'toggle auto-restore')
chat uiUP on 127.0.0.1:3000
gpu1x H100 NVL
diskmodels at /mnt/nvme/models
httpsnot exposed (localhost + SSH tunnel only)
ufwactive

----------------------------------------

Caleta appliance - choose an action

←↓↑→ navigate · enter submit · here: click an action to run it

A simulation of the real on-box menu. The real one needs no SSH client: it runs in the Azure portal serial console.

  • Private open-weight AI inside your Azure tenant. Nothing leaves your subscription.
  • From deploy to first answer in minutes, not days or weeks.
  • Azure Spot capacity with live ranking of the cheapest, most stable regions.
  • Up and running in as little as 5 minutes with our quick start.
  • Park it overnight: only the OS disk keeps costing money.
  • Honest numbers: where we quote a speed, we measured it on this exact image.

Three ways we can help

Run AI on your own terms. Spend less doing it.

The Private AI Appliance

A ready-to-run Azure Marketplace image: open-weight AI models, a private chat interface, and an honest model catalogue. One deployment builds the network, the firewall and the box. Running in minutes, from a browser, and nothing leaves your subscription.

Managed Private AI

The appliance, engineered into a resilient multi-region private AI service: designed around your workloads and your network, priced from measured eviction data, built inside your tenant. You own everything we deploy.

Azure Cost Review

A free, read-only scan of your Azure spend, sized savings in pounds, not a dashboard to decode. If the findings justify it, we offer a verified review and hands-on implementation.

What We Do

Private AI in your own Azure subscription, expert consulting, and hands-on infrastructure services that optimise your IT spend.

Private AI Appliance

A ready-to-run Azure Marketplace image that serves open-weight AI models privately in your own subscription. Your models, your tenant, your data.

  • One deployment builds network, firewall and box
  • NVIDIA, AMD, CPU and Arm: one image
  • Private chat UI and OpenAI-compatible API
  • Honest, dated benchmarks on real Azure hardware
Learn more

Azure Consulting & DevOps

Hands-on Azure architecture and DevOps from a certified Solutions Architect. We build, automate, and deliver - not just advise.

  • Azure architecture & review
  • CI/CD & DevOps pipelines
  • Infrastructure as Code
  • AI-augmented delivery
Learn more

FinOps & Cost Optimisation

Identify wasted cloud spend and optimise your Azure environment. Most businesses save 20-40% after optimisation.

  • Free savings snapshot
  • Self-service & Blueprint options
  • Full implementation available
  • Ongoing monthly support
Learn more

Cloud Migration

Migrate your on-premises infrastructure to Azure with expert guidance. Certified Azure Solutions Architect on hand.

  • Migration assessment
  • Architecture design
  • Lift-and-shift or modernise
  • Minimal downtime
Learn more

Colocation Services

Rack space in the Slough data centre corridor with bundled Smart Hands support. One relationship for facilities and service.

  • Slough DC corridor access
  • Bundled Smart Hands support
  • DC-to-DC migrations
  • Hybrid cloud connectivity
Learn more

Smart Hands & Remote Hands

Expert data centre support in the Slough corridor. Out-of-hours and emergency cover from engineers who troubleshoot, not just follow runbooks.

  • Out-of-hours cover
  • Rack & stack, hardware swaps
  • Network troubleshooting
  • Same engineer every time
Learn more

Data Shuttle

Move terabytes in hours, not weeks. White-glove data transfer to Azure, AWS, or any data centre.

  • 10Gbit+ transfer speeds
  • AES-256 encryption
  • 24-48 hour turnaround
  • Full chain of custody
Learn more

Why Choose Caleta.io?

Local Presence

Based in Slough, the heart of the UK data centre corridor. Face-to-face meetings available.

Microsoft Certified

Azure Solutions Architect certified. Microsoft Partner network member.

No Obligation

No strings attached. If we don't find savings, we'll tell you.

30+ Years Combined

Deep expertise in enterprise IT, cloud infrastructure, and cost optimisation.

Trusted by Big 4 professional services firms, FTSE 100 energy companies, and global technology providers.

Our engagements regularly turn into 2+ year partnerships. Clients stay because we deliver.