AI Access

Ollama

Run supported open models on your own computer through desktop apps, a command line, or a local API.

See how this access route works, what it asks you to set up, and which current terms to check before using it.

Open provider Back to AI Access

May fit if

You want a local model runtime

Local experiments, private files that should stay on-device, and learning how models run.

Access observed

Free local use

Ollama lists a $0 Free plan. Local models run on your own hardware without a usage quota; the Free cloud route includes starter credits for selected models.

What you will need

Installation and suitable hardware

Install Ollama on supported hardware, choose a model that fits the machine, and decide whether each request should stay local or use an optional cloud model.

Local and cloud routes now sit in one product

Ollama runs supported models on your hardware through apps, a CLI, and a local API. Its Free plan also includes starter credits for selected cloud models, which use a remote data path.

Limits and charges

Local hardware and cloud credits differ

Local performance depends on your computer and model size. Free cloud credits cover a smaller starter-model set and reset monthly; purchased credits unlock other cloud models.

Your data and account

Local prompts stay local only on the local route

Ollama says it does not receive local prompts or responses. Cloud models process content remotely, so confirm the active model and route before using private material.

Before relying on it

Check the route, model size, and exposed API

Local execution does not automatically make every connected workflow private or safe. Check model files, integrations, exposed ports, cloud routing, and hardware limits. Confirm hardware fit, disk use, model source, network exposure, connected applications, and whether the selected model is local or cloud-hosted.

Check current information