Ollama (Local Models)
Run local models like Llama, Mistral, and Gemma on your own hardware
Last updated: July 6, 2026
Overview
Ollama runs LLMs locally, using your computer's hardware instead of the cloud. It has its own page at Settings → Ollama.
You can chat with local models without an internet connection, and since nothing leaves your machine, it's more private than cloud providers. It's also free to use — Ollama is available on all plans, including Starter.
You need a decent internet connection to download models, plus enough disk space to store them. Local model speed depends entirely on your hardware — a capable CPU/GPU and enough RAM make a big difference.
Setting Up Ollama
-
Install and run Ollama. If Ollama isn't detected, the page shows Please Start / Install Ollama with two buttons: Configure (point Slashspace at a non-default Ollama address) and Install Ollama (opens the download page at ollama.com). Install it, start it, and come back.
-
Once Ollama is running, the page shows an Installed badge with the Ollama version.
-
Find a model ID. Browse ollama.com/search for the model you want and copy its ID.
-
Download it. Back in Settings → Ollama, paste the model ID (e.g.
gemma3:3b) into the input (placeholder: "ex: mistral") and click Download Model. Progress streams in as the model downloads. Once it finishes, the model appears in the list and is available across Slashspace.
The Search Models on Ollama link below the list takes you straight to the model catalog.
Managing Models
- Models that are currently loaded show a green Running indicator and sort to the top of the list.
- Each model row has a Delete button to remove it and free up disk space.
Configuration
Click the gear button next to the Ollama badge to open the Ollama Configuration dialog. It has a single Hostname field — the URL of your Ollama instance, http://localhost:11434 by default (a "use the default" shortcut fills it in). Click Save Configuration to apply.
Point this at another address if you run Ollama on a different port or on another machine on your network.
Troubleshooting
- Verify Ollama is running at
http://localhost:11434 - Check your firewall settings
- Restart Ollama if needed
More help in Troubleshooting & Support.
Use cases
- Fully offline work: download a model once, then chat, take notes, and build canvases on a plane or anywhere without internet.
- Private material: keep sensitive documents and conversations entirely on your machine — nothing is sent to any cloud provider.
- Zero-cost experimentation: try prompts, personas, and node workflows freely on the Starter plan without spending credits or API budget.
- A home LLM server: run Ollama on a beefy desktop and point your laptop's Hostname setting at it over the local network.