Pricing
Pay for the fleet, never by the token.
Your hardware does the work, so nothing here is metered per prompt or per token. Plans follow how many machines you run and how you like to run them.
Home
Make my GPU useful.
One machine, one GPU. No card and no account needed.
A great way to turn a desktop or a Mac into your own private AI endpoint.
Home includes
- Models on your own GPU or Mac
- The curated catalog, verified downloads and fit checks
- Chat everywhere, plus embeddings, reranking, structured output, tool calling and vision on Linux
- Your own GGUF models, imported and qualified on Linux
- An OpenAI-compatible API with streaming
- The CLI, the local dashboard and
nodeau doctor - Your machine in your account, visible from any browser
Home Pro
Make all my GPUs work together.
For one person with more than one machine. Cancel any time.
Your desktop, the server in the cupboard and the laptop that wants to reach both, as one fleet.
Everything in Home, and
- Up to three machines in one fleet
- Up to two GPUs in each machine, one model per card or one model across both
- A model on every GPU in your fleet, all running at once
- Batch inference, with several workers at once
- Run your fleet from any browser: start and stop models, drain machines, read logs, recover work
- Workloads that come back up on another of your machines by themselves, from a copy of the model kept ready there
- Set your fleet's limits from the web
- Upgrade the whole fleet from one place, one machine at a time
- Standard support
Business
Our organisation's private AI fleet.
Tailored to your fleet and priced per organisation.
For teams running models across the workstations and servers they already own.
Everything in Home Pro, and
- As many machines and GPUs as you run
- Business support, and a direct line to the people building Nodeau
- A hand planning your setup, policies and rollout
- A say in what we build next
Planned for Business
- Members with roles, single sign-on and scoped keys for scripts. See the roadmap
Side by side
The free plan is a whole product on its own. The paid ones add machines and more ways to run them.
| What you get | Home | Home Pro | Business |
|---|---|---|---|
| Run models | |||
| Models on your own GPU or Mac | Included | Included | Included |
| Curated catalog, verified downloads, fit checks | Included | Included | Included |
| Embeddings, reranking and vision, on Linux or a Mac | Included | Included | Included |
| Tool calling and structured output, on Linux | Included | Included | Included |
| Bring your own GGUF model, on Linux | Included | Included | Included |
| OpenAI-compatible API, CLI and local dashboard | Included | Included | Included |
| Scheduling modes and power limits for NVIDIA GPUs | Included | Included | Included |
| Accept a fit risk yourself, per model | Included | Included | Included |
| Scale | |||
| Machines | 1 | Up to 3 | As many as you run |
| GPUs in one machine | 1 | Up to 2 | As many as you run |
| Models running at once | One per GPU | One per GPU | One per GPU |
| Batch inference | Starts with Home Pro | Included | Included |
| Your account | |||
| See your machines from any browser | Included | Included | Included |
| Usage in accelerator-hours and a record of changes | Included | Included | Included |
| Entitlements for machines with no internet | Included | Included | Included |
| Remote dashboard and management | Starts with Home Pro | Included | Included |
| Set your fleet's limits from the web | Starts with Home Pro | Included | Included |
| Fleet upgrades from one place | Starts with Home Pro | Included | Included |
| Support | Community | Standard | Business |
Everything in this table works in the current release. Your plan arrives as a signed entitlement that each machine checks on its own, so your models keep serving when a machine is offline. What we're building next lives on the roadmap.
Why there's no per-token price
Your GPU does the work
The compute is hardware you already paid for. We'd rather charge for making it useful than for your own electricity.
The value is running it
Working out what fits, keeping it running and knowing what's where. That grows with machines, so that's what plans follow.
Costs you can plan
Owned hardware has a known capacity and a known cost. For steady work that's a comfortable shape to budget around.
The economics depend on how busy your machines are. Hardware that runs work most of the day gets the most out of owning it.
Start on the machine you already have.
Home is free, needs no account, and Nodeau tells you straight away what your machine can run.