Ollama Desktop

Ollama support is for Aori Desktop. Aori connects to a local Ollama server that you install and run on your Mac, Windows PC, or Linux machine.

Aori does not bundle local models, manage model downloads for you, or provide native mobile local inference in this launch. Android and iOS users should choose a cloud provider such as Groq, OpenAI, Anthropic, Gemini, DeepSeek, or OpenRouter.

Note

Ollama can keep model prompts on your desktop for requests handled entirely by Ollama. Data may still leave your device if you use cloud Skills, web search, messaging channels, support emails, website analytics, or a cloud AI provider for another request.

Requirements

• Aori Desktop for macOS, Windows, or Linux.
• Ollama installed from ollama.com.
• At least one downloaded model.
• Enough memory and CPU/GPU capacity for the model you choose.

Install and Prepare Ollama

1. Download and install Ollama.
2. Launch Ollama and confirm it is running.
3. Pull a model:

ollama pull llama3.2

4. Confirm the model is available:

ollama list

Connect Aori to Ollama

1. Open Aori Desktop.
2. Go to Settings -> AI Models.
3. Choose Ollama.
4. Keep the default local endpoint unless you changed it:

http://localhost:11434

5. Select one of your installed models.

What to Expect

Local models can be slower and less capable than hosted models, especially for tool-heavy tasks. If Aori says a model does not support a required capability, switch to another local model or use a cloud provider for that task.

Good starting points:

Model	Best for
llama3.2	General chat and lightweight tasks
mistral	Instruction-following and drafts
llama3.1:8b	Better reasoning on capable machines

For first-time Aori setup, Groq is usually the fastest path because it has a free tier and does not require local model setup.