Your models. Your cloud. Your knowledge. One private AI platform.
Production AI inference on infrastructure you control — with your organisation's governed knowledge built in. OpenAI-compatible. Region-locked. Ready in minutes.
No Kubernetes expertise required. No cloud accounts to configure. CaseDesk handles infrastructure so your team can focus on the product.
1
Answer five business questions
Use case, team size, region, and compliance needs. No cloud account required at this stage.
2
Review your recommended deployment
CaseDesk selects the model class, GPU tier, and region. You see what you will pay before committing.
3
Your endpoint is live
A dedicated inference endpoint, OpenAI-compatible. Drop the URL into Cursor, Continue.dev, or your own app.
CaseDesk Dedicated
Your model. Your endpoint. Your region.
Private AI deployment with your organisation's governed knowledge built in — running on infrastructure you control.
No shared inference queues.
No external model providers.
No data leaving your chosen region.
You retain full control while CaseDesk manages deployment, upgrades, monitoring, and operations.
🔒
Data stays in your region
Your prompts and completions never leave the UK, EU, or US region you choose. CaseDesk does not log or store inference traffic.
🎯
One endpoint per team
No shared inference queues. Your deployment runs in its own namespace on a dedicated GPU cluster. Standard tier uses a shared node pool. Enterprise tier adds hardware isolation.
💸
Scale to zero when idle
Deployments scale to zero after a configurable idle period. Your endpoint stays available on a flat monthly subscription - no per-request charges, no usage spikes.
🔧
OpenAI-compatible API
Every endpoint implements the OpenAI REST API. Swap the base URL in your existing tools - no code changes required.
Integrations
Works with the tools you already use
Any OpenAI-compatible client connects without modification.
Open WebUI
Chat interface with file upload, RAG, and tool support
Cursor
AI-native editor, drop in as a custom model
Continue.dev
VS Code and JetBrains AI coding assistant
LangChain
Chain-based AI applications and agents
OpenAI SDK
Python and TypeScript SDKs, zero changes needed
n8n
Workflow automation with AI nodes
Flowise
Low-code LLM app builder with visual flows
Ollama API
Compatible with Ollama-format API clients
Pricing
Simple monthly pricing
One flat monthly fee. Web search, model routing, and your organisation's knowledge included. UK, EU, and US regions available.
Dedicated hardware node, no other customer workloads on the same GPU. Required for NHS Trusts, financial services, and regulated organisations. Quoted after Architecture Review.
Every new account gets a sandbox deployment automatically. Shared GPU, UK region, scale to zero after 1 hour. Use it to try a model before moving to a dedicated deployment.
CaseDesk Enterprise provides dedicated node pools, hardware isolation, and infrastructure designed to support NHS DSPT and GDPR compliance. Your data and inference traffic never leave the UK region.
The sales process includes Discovery, Technical Workshop, Architecture Review, and Security Review before any data touches production infrastructure.
Dedicated GPU node - no shared hardware
UK data residency, DSPT-ready
SSO and RBAC
Private networking options
SLA and support contract
Architecture Review included
We start with a 30-minute discovery call. No slide decks - just the technical questions that matter for your compliance requirements.
Your deployment runs in its own namespace on a regional GPU cluster. You get a dedicated inference endpoint - no other team's traffic shares your model. Standard tier uses a shared node pool. Enterprise tier adds hardware isolation: your model runs on a physical GPU that no other customer touches.
Your inference requests and completions never leave your chosen region, and CaseDesk does not use them to train models. CaseDesk runs open-source models on infrastructure it controls in UK or EU data centres, with OpenAI-compatible APIs. Optional web search and page-reading tools may contact external providers when enabled.
Answer five questions in the Deploy Wizard and your endpoint is live in minutes. No Kubernetes knowledge required. CaseDesk handles provisioning, scaling, and maintenance.
Your inference requests and outputs stay within the region you choose: UK (AWS eu-west-2), EU (Azure westeurope), or US (GCP us-east1). CaseDesk does not log or store your prompts or completions.
Yes. Every CaseDesk endpoint exposes the OpenAI REST API. Swap the base URL and your tools - Cursor, Continue.dev, LangChain, n8n, or any OpenAI SDK - work without code changes.
Yes. CaseDesk Enterprise tier provides dedicated hardware isolation, UK data residency, and infrastructure designed to support NHS DSPT compliance. Patient data and clinical information never leave the UK region.
Deployments scale to zero after a configurable idle period. Your monthly subscription continues regardless - you are paying for the private endpoint, data residency, and tooling, not for individual requests. Scale-up from zero takes 60-120 seconds.
Yes. You can redeploy to a different model at any time from the deployment detail page. Your endpoint URL stays the same.
Give your team a trusted AI platform.
Start with the sandbox. Move to a dedicated deployment when you are ready.
