Choosing the Right Plan
Managed CaseDesk is a flat monthly subscription. Choose the plan that matches your model size and concurrency requirements.
Plans
| Plan | Monthly price | Model size | Runtime | Max concurrent users |
|---|---|---|---|---|
| Starter | £199 / €229 / $239 | 1–8B | Ollama (sequential) | 1 |
| Team | £449 / €499 / $519 | 9–20B | vLLM | 20 |
| Advanced | £1,199 / €1,349 / $1,399 | 21–70B | vLLM | 50 |
| Enterprise Isolation | Custom | Any | Dedicated hardware node | Custom |
Prices are per deployment, billed monthly. No GPU-hour charges. No per-seat charges.
Contact sales for Enterprise Isolation pricing.
Concurrency
Starter uses Ollama, which processes requests sequentially. It is suited to single-user or low-concurrency workloads.
Team and Advanced use vLLM continuous batching. Multiple users can generate responses simultaneously from a single pod:
- Team — up to 20 concurrent users generating at once
- Advanced — up to 50 concurrent users generating at once
The flat monthly price covers all concurrent users with no additional per-seat charges.
Choosing a model
Within each plan, you choose a model from the catalogue at deployment time. Guidance by use case:
| Use case | Recommended model range |
|---|---|
| Assistants, chatbots, customer support | 1–8B (Starter) |
| Complex reasoning, agents, document analysis | 9–20B (Team) |
| Large-context tasks, long documents, legal/medical analysis | 21–70B (Advanced) |
:::tip Evaluating hosted alternatives? If you're comparing CaseDesk against GPU cloud or serverless inference providers, see the comparison pages for a side-by-side breakdown on cost, data ownership, and control. :::