Documentation

Everything you need to get started with Saga.

Quick Start

1. Download Saga

Get Saga for your platform from our download page.

2. Install Ollama (for local AI)

Saga will prompt you to install Ollama on first launch, or you can install it manually:

curl -fsSL https://ollama.ai/install.sh | sh

3. Download a model

Browse our model catalog or pull one directly:

ollama pull qwen2.5:7b

4. Start chatting

Open Saga, select your model from the Local AI tab, and start your first conversation. Your data stays on your machine.

FAQ

Do I need an internet connection?

No. Local AI runs completely offline. You only need internet to download models or use cloud integrations (GPT, Claude).

What are the system requirements?

For basic models (3B-7B): 8GB RAM, any modern CPU. For larger models (14B+): 16GB+ RAM, dedicated GPU recommended.

Is my data sent anywhere?

With local models: no, everything stays on your machine. With cloud integrations: requests go to the respective provider (OpenAI, Anthropic), but Saga logs all outbound requests so you can see exactly what's sent.

Can't find what you're looking for? Contact support