Choose where your AI runs.
Noema keeps the route visible. Local execution and cloud access serve different tasks, and you can choose between them.
| Route | Useful when | What to consider |
|---|---|---|
| Your Browser | You want model processing on the device you are using | Graphics support, available memory and model download size |
| Your Machine | You already run a local model server | Server availability, browser access and the model’s capabilities |
| Noema Router | You want automatic cloud selection | Zero-retention Auto routing, ZKR, wallet-funded credits |
| A named cloud model | You want a specific model | Price, input support, context and policy labels |
| Noema Network | Future independent confidential compute | Planned, not available yet |
A local model does not make every tool offline
The initial model download needs a connection. Optional web search and external tools can also use the network. Consider both the model route and the tools enabled for your request.
Keep model choices current
Use the models shown in Noema. Available models, prices and policies can change. A model supporting an input or output type does not mean every possible workflow is connected in the workspace yet.